Special Issue: Autonomic ComputingandApplicationsforAutonomous Systems Guest Editor: Weiping Zhang Editorial Boards Informatica is a journal primarily covering intelligent systems in the European computer science, informatics and cognitive com­munity; scientifc and educational as well as technical, commer­cial and industrial. Its basic aim is to enhance communications between different European structures on the basis of equal rights and international refereeing. It publishes scientifc papers ac-ceptedbyat leasttwo referees outsidethe author’s country.Inad­dition, it contains information about conferences, opinions, criti­calexaminationsofexisting publicationsandnews. Finally,major practical achievements and innovations in the computer and infor­mation industry are presented through commercial publications as well as through independent evaluations. Editing and refereeing are distributed. Each editor from the Editorial Board can conduct the refereeing process by appointing two new referees or referees from the Board of Referees or Edi­torial Board. Referees should not be from the author’s country. If new referees are appointed, their names will appear in the list of referees. Each paper bears the name of the editor who appointed the referees. Each editor can propose new members for the Edi­torial Board or referees. Editors and referees inactive for a longer period can be automatically replaced. Changes in the Editorial Board are confrmed by the Executive Editors. The coordination necessary is made through the ExecutiveEdi-tors whoexamine the reviews, sort the accepted articlesand main­tain appropriate international distribution. The Executive Board is appointed by the Society Informatika. Informatica is partially supported by the Slovenian Ministry of Higher Education, Sci­ence andTechnology. Each author is guaranteed to receive the reviews of his article. When accepted, publication in Informatica is guaranteed in less than one year after the Executive Editors receive the corrected version of the article. Executive Editor – Editor in Chief Matjaž Gams Jamova 39, 1000 Ljubljana, Slovenia Phone: +38614773 900,Fax: +38612519385 matjaz.gams@ijs.si http://dis.ijs.si/mezi/matjaz.html Editor Emeritus AntonP. Železnikar Volariˇceva 8, Ljubljana, Slovenia s51em@lea.hamradio.si http://lea.hamradio.si/~s51em/ Executive Associate Editor -Deputy Managing Editor Mitja Luštrek, Jožef Stefan Institute mitja.lustrek@ijs.si Executive Associate Editor -Technical Editor DragoTorkar, Jožef Stefan Institute Jamova 39, 1000 Ljubljana, Slovenia Phone: +38614773 900,Fax: +38612519385 drago.torkar@ijs.si Contact Associate Editors Europe, Africa: Matjaz Gams N. and S. America: Shahram Rahimi Asia, Australia: Ling Feng Overview papers: Maria Ganzha,Wies awPaw owski, Aleksander Denisiuk Editorial Board Juan Carlos Augusto (Argentina) Vladimir Batagelj (Slovenia) Francesco Bergadano (Italy) Marco Botta (Italy) Pavel Brazdil (Portugal) Andrej Brodnik (Slovenia) Ivan Bruha (Canada) Wray Buntine (Finland) Zhihua Cui (China) Aleksander Denisiuk (Poland) Hubert L. Dreyfus (USA) Jozo Dujmovi´ c (USA) Johann Eder (Austria) George Eleftherakis (Greece) Ling Feng (China) VladimirA.Fomichov (Russia) Maria Ganzha (Poland) Sumit Goyal (India) Marjan Gušev (Macedonia) N. Jaisankar (India) Dariusz Jacek Jakczak (Poland) Dimitris Kanellopoulos (Greece) Samee Ullah Khan (USA) Hiroaki Kitano (Japan) IgorKononenko (Slovenia) MiroslavKubat (USA) Ante Lauc (Croatia) Jadran Lenarciˇˇ c (Slovenia) Shiguo Lian (China) Suzana Loskovska (Macedonia) Ramon L. de Mantaras (Spain) Natividad Martínez Madrid (Germany) Sando Martinciˇ´c (Croatia) c-Ipiši´Angelo Montanari (Italy) Pavol Návrat (Slovakia) Jerzy R. Nawrocki (Poland) Nadia Nedjah (Brasil) Franc Novak (Slovenia) MarcinPaprzycki (USA/Poland) Wies awPaw owski (Poland) Ivana Podnar Žarko (Croatia) Karl H. Pribram (USA) Luc De Raedt (Belgium) Shahram Rahimi (USA) Dejan Rakovi´ c (Serbia) Jean Ramaekers (Belgium) Wilhelm Rossak (Germany) Ivan Rozman (Slovenia) Sugata Sanyal (India) Walter Schempp (Germany) Johannes Schwinn (Germany) Zhongzhi Shi (China) Oliviero Stock (Italy) RobertTrappl (Austria) TerryWinograd (USA) Stefan Wrobel (Germany) Konrad Wrona (France) XindongWu (USA) Yudong Zhang (China) Rushan Ziatdinov (Russia&Turkey) Report from IJCAI 2019, Top AI Conference for 50 years Which Crisis is Coming First – of AI or of World Economy? Matjaž Gams Jožef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia E-mail: matjaz.gams@ijs.si Editorial Introduction This year IJCAI [1] -"International Joint Conference on Artificial Intelligence" -celebrates half a century of continuous conferencing as the world AI's most important global event. Over the last ten years, the number of submissions has grown steadily, by more than 30% in the last two years alone, and for 2019 it approached 5000 with 2700 committee members (Figure 1). The acceptance rate this year was less than 18%, one of the lowest ever, leading to 850 papers presented at the main conference, accompanied by three days of workshops. The papers from China (327, 38%) toppled the papers from EU (152) and USA (169) combined, while other countries followed with Australia (37), India (20), Japan (18) and areas like Eastern Europe or Africa far behind. The shift in recent years has been enormous, not only in terms of the number of papers, but also in terms of the number of new applications and the overall focus of countries and resources involved. The conference costs for one speaker, for example, amounted to several thousand euros. In terms of rewards, lifetime achievements and invited lectures, USA and EU still dominated due to inertia since the number of senior AI researches in Asia has only recently started to increase. Besides Program Committee Chair Sarit Kraus, there were Tutorials Chairs, Workshop Chairs, Demo Chairs, Doctoral Consortium Chair, Robot Exhibition Chairs, Video Competition Chairs, Survey Chairs, and Chairs for Sister Conference Best Papers, Journal Track, Special Track on Understanding Intelligence and Human-level AI in the New Machine Learning era, Special Track on AI for Improving Human-Well Being. The distribution of IJCAI papers in 2019 by area, submitted and accepted is presented at Figure 2, some competitions in Figure 3. 2 Achievements and dilemmas IJCAI is not only a conference, it is an annual presentation of the world's AI best and brightest and most relevant events, e.g. the meeting of world AI societies. Unfortunately, there are glitches, and this year the presence of national AI representatives was more than sporadic. Hopefully in 2020 the organizer will send an invitation to all national organizations in time. Another idea -as this is an informal organization, we will draw up a list of all national AI societies and submit an invitation in time. Better two invitations than none. In the world of over-aggressive web and other media advertising and social media, activities of scientific societies are somehow overshadowed. For example, while AI funds around the world are growing rapidly with an EU annual increase of around 50%, European AI societies (EurAI) have not increased their memberships, and no new European AI society joined in 2019. When asked why AI societies are not more aggressive in trying to attract more societies and members, the reply was that this can hardly be expected from a scientific organization -e.g. to use commercial advertising methods. Maybe, or maybe not and what is needed are strong vision and determination. C:\Users\mezi\Documents\Potovanja\IJCAI2019\n submissions.jpg 2007 2009 2011 2013 2015 2016 2017 2018 2019 Figure 1: Submissions to IJCAIs. Figure 2: IJCAI 2019 submitted and accepted papers by area. While IJCAI 2019 as a whole was undoubtedly a great success, some issues were observed in specific areas. For example, "AI in industry" presented several important AI applications. To name one, Xiaowei, a Chinese mobile platform with several AI modules, has around one billion users (Huaowei around 3 billion) and is constantly introducing new AI services. At the 2019 IJCAI conference they presented an AI assistant. When fully implemented and reaching all users, it will bypass Siri (0,5 billion), Google Home (0,4) and Alexa (0,1) users as current leading voice-controlled AI assistants. In the AI demonstration section, a fast growing and flourishing area of IJCAI, several systems deserved and often got world­wide appreciations. For example, one system improved distribution and placement of Uber drivers, enhancing the individual driver’s gain and decreasing the user’s waiting time. One rewarded demo presented an automatic creation of assistants from websites and the other fair use of workforce. The industry AI award was given to a Microsoft team for an application of reinforcement learning to personalize news (28% increase in adaptation) and games (40%). These huge improvements were obtained by relatively small modifications of the previous systems, a lesion to be remembered. But surprisingly, the participation in the lecture hall was less than average. The explanation at hand M. Gams is that most of the conference attendees were researchers listening to academic presentations that took place simultaneously in several parallel sections. The gap between academia and industry was highlighted once again. Researchers receive funding and fame according to academic criteria and it is not of great importance if their ideas find ways to actually help people and increase profits. The fusion of real-live applications and academia at IJCAI was courageously attempted in many respects, such as the competition for care of the elderly instead of robot soccer (Figure 4). The two major advances in 2019 compared to 2018 were probably the increase in massive AI applications, and secondly, new research orientations. The fact that the former was somehow accepted as an obvious fact is not very helpful for AI growth and fame, where an AI program beating humans in a particular game obtains overall attention, while major AI applications hardly ever. But it is precisely the dozens of industry presentations, demos, workshops, competitions such as the elderly-care competition (Figure 4) and practical AI presentations, often related to a particular branch such as robots, that have most impressed an impartial AI observer in 2019. AI is in the intense phase of transforming human society into an advanced, incremental, optimized and multi-objective civilization providing better foundations for long-term sustainable growth. As usual, there were hundreds of incremental algorithm improvements, be it random forest, boosting or deep neural networks. In particular, the deep learning, where a random forest algorithm is placed instead of a neuron in a network, had shown quite important improvements. The problem, which is consistent with the principle of multiple knowledge [2], is that it quickly loses diversity with additional levels since the random forest consists of more or less all possible decision trees. Therefore, despite some interesting results the original idea of adding another algorithm such as RF instead of a neuron is still in progress. Overall, this incremental progress is quite impressive as AI is used for hundreds of trillions of decisions a day, and a few percent better decisions mean a lot in real life. C:\Users\mezi\Documents\Potovanja\IJCAI2019\angry birds.jpg C:\Users\mezi\Documents\Potovanja\IJCAI2019\competitions.jpg Figure 3: Competitions at IJCAI 2019. C:\Users\mezi\Documents\Potovanja\IJCAI2019\n elderly comp.jpg Figure 4: Elderly-care competition instructions. The focus change from academic to real-life was noticeable at IJCAI 2019. AI directions As far as the new research orientation is concerned, at IJCAI 2019 it was carried out in a rather precise way, although not by all the invited speakers. But several lectures merged into a new research paradigm that was best presented at the conference last afternoon by Broeck, Domingos, and Shoham. It is not a switch from Decision Tress (DTs) to Deep Neural Networks (DNNs) because DTs (or decision graphs) are as accurate as DNNs in some areas and also provide explanations and understanding that are unprecedented for DNNs, regardless of the DNN advances outperforming any artificial system in several areas and often the best people. Nor is it very likely that DNNs will soon become truly intelligent, as Figure 5 shows -a boy standing on a chair hoping in vain to see the stars better with a telescope. It's also not a dilemma whether to use AI systems like GPT-2, HAIM, Grover or "Not Jordan Peterson" (the last one removed due to lawsuits) -why not use them for fun and get acquainted with the power of SOTA AI. AI is a technology and like any technology it can be used for good or bad. At present, AI it is the one that contributes most to human progress, with applications ranging from robotics to web services and autonomous cars. According to practical statistics, a Tesla car, for example, is nine times safer in the autonomous driving mode than an average classical car. The progress can also be seen in the services openly available on the net -for example, a few years ago nobody could create a system like Not Jordan Peterson -fluently speaking input text that does not differ from the speech of original author. The actual question / dilemma, according to the IJCAI presenters is the following: Are we on the path to developing truly intelligent systems or only AI applications capable of playing excellent chess, for example, where the specific algorithmic solutions are dedicated and successful only in a certain area, without explanation and without the impression that something inside resembles a real human intelligence? Since the attempts to solve the Turing test remain as unsuccessful as ever while the computer and AI progress is continuing with the exponential speed, something soberer seems to be hidden in this perplexity. However, there are the good old strong AIers who claim that we are on the way to true intelligence and we just need to be patient a little more. And anyway, who says that intelligent AI systems need to have human-like intelligence to perform well, because airplanes fly in contrast to birds, while the direct applications of bird-like flight patterns is counterproductive. Why should an autonomous car write or understand sophisticated poems about ethics, mortality or love? Furthermore –consider the moral dilemma of autonomous cars: Who should a car hit -a child or a grandma, if it cannot stop in time? In practice, this is statistically an irrelevant question, since such a situation practically never occurs in an individual's life. Second, in some countries, such as Germany, there is a law that forbids taking preferences based on age -a car that prefers to run over grandma would therefore be illegal and subject to legal consequences. And thirdly, why is this dilemma imposed on the scientific engineering community, where 99% of activities are aimed at developing technical solutions that enable high-quality driving in all possible real circumstances from weather conditions to the reactions of other road users? Should we concentrate our energy on a hypothetical situation or rather design better systems to save thousands of lives Figure 5: Probably, current deep neural networks, alike other existing AI approaches, will never achieve true intelligence on their own. New strategies are needed. each year? It should also be noted that two decades ago autonomous cars were more of a joke than something people believed in -so amazing is the AI progress! So why is a large part of all discussions dedicated to a philosophical dilemma of whom should a car strike first in more or less artificial situations? Why does a Tesla incident caused by an AI mistake attract a lot of media attention and the corresponding ten situations in which it avoided a crash with superhuman reflexes practically none at all? Not to mention that there are several benchmark domains where AI progress can be demonstrated explicitly in a scientific, repeatable and measurable way. In reality, for a large part of the AI applications, no real human-level intelligence is required and the engineering AI already offers significant improvements. For example, DNNs in the ImageNet benchmark visual recognition test improved their accuracy from 71% to 93% from 2011 to 2014 and from 2014 to 2018 to 98%. To name some of the most important AI achievements in recent years: 2017 -skin cancer, poker; 2018 – SQUAD1.0, Chinese-English translation, Dota2, prostate cancer; 2019 -SQUAD2.0, Starcraft. DNN combined with reinforcement learning enabled a big step ahead. For example, Google's DeepMind played magnificently 50 Atari games. An example game would be hitting the ball from the bottom of the screen upwards to hit objects in several rows at the top of the screen to score points. But to demonstrate the strange nature of some recent AI achievements, when the racket was moved up a few pixels, the performance deteriorated significantly, which is highly unlikely for humans. Another example -when it was investigated how DNNs learned to distinguish cats from dogs, it turned out that their decisions were based on a small number of pixels and when these pixels were changed, with humans still clearly distinguishing between different animals, DNNs failed. These days, researchers are developing algorithms that compete in the search for the minimum number of pixels that need to be changed to mislead DNNs. Several experiments of this kind showed that DNNs learn significantly different than people do, that their knowledge is not general, but highly specialized and therefore brittle. The expectation that DNNs, like other existing AI technologies, will achieve true intelligence is rather an attempt shown in Figure 5, according to several distinguished AI researchers. Figure 6: The issue of decreasing number of women in computing is a relevant and sound one, but the idea to use other criteria instead of research excellence for scientific publications is a threat to science. Figure 7: New strategic AI research direction on a path towards true intelligence, proposed by Shoham. AI dangers Among the dangers for AI, another one lurks in the shadows -the penetration of ideology, which is currently happening in all spheres of human activity, be it mass media or science. Indeed, in some AI areas like superintelligence [3,4], scientific objectivity, i.e. novelty, should not be the only criterion. A hostile superintelligence could harm human civilization, and therefore such attempts should be treated with appropriate supervision and caution[5]. For AI conferences and journals, the standard penetration of ideology is not a major problem, as all these activities are based on anonymous refereeing where authors based on personal characteristics such as position, country of origin, wealth, skin color, gender, age or similar cannot be preferred. In IJCAI 2019, however, there were some attempts to modify this standard approach. They all demanded a "fair" share of a certain section of the population -note that this paper will strictly avoid naming such criteria. Science and ideology do not walk along well, and it is disturbing for any true scientist to observe the growth of ideology in recent years. It might be too early to warn that virtually all civilizations have saturated and collapsed with the growth of harmful ideologies -harmful in the way they collided with the production of vital goods. In Western civilization, the negative effects of overwhelming neo-liberal globalization are quite obvious, from the overburdening of the planet by transporting industries to less developed countries with cheap labor and thus overloading the global environment, to the structure of important positions based on political orientations and personal categories such as gender or skin color, and not primarily on the ability to work well. Be what it may, the declining number of women in IT technology (Figure 6) is indeed problematic. As a rule of thumb, at least 20% to 30% of the members of the opposite sex are needed to achieve good group performance and according to statistics this is already hard to achieve in some teams. In comparison -for boards of directors, CEOs or ministries, many Western countries require 40% of represents another extreme, and would be indeed unfortunate if such criteria were introduced into science, as softly advocated. But as long as the refereeing remains anonymous, there is no direct way for ideology to corrupt science too much. At IJCAI, there wasn't much talk about media IT giants like Google or Facebook, but the way they affect human society to become more polarized, hostile and less open is becoming more and more evident. As an indicative example, suppose one observes a YouTube clip claiming that the Earth is flat. The recommendation system observes the area of interest and recommends more videos that the Earth is flat. Soon all one person will get are videos confirming the wrong belief. There are other stray effects related to the IT giants. Perhaps we should not discuss the penetration of ideology into say Google by deliberately providing the objective algorithms with false data in order to learn to eliminate politically undesirable persons from the media light (e.g. name modifications) as it is probably not a major phenomenon. However, the IT giants with all their positive novelties have also the sinister monopolistic side, influencing elections, and making people more stupid, as the Flynn effect shows with centuries of progress and a decline in the last decade. At the same time, when used properly, fair and without ideological twists [5], AI will continue significantly improving the quality of the Web. For example, violent videos are removed from YouTube much faster and more efficiently than before with YouTube AI guards. From time to time there are still some failures, such as the elimination of robotic wars with the argument that cruelty to animals exists, but overall the improvement is indeed significant. 5 Conclusion and discussion To summarize: - Even without true AI, the incremental AI progress with rather engineering solutions already offers great improvements, and all attempts to discredit them or to shift the discussion outside the scientific, engineering Figure 8: The last page of the IJCAI conference. solutions are doomed to fail due to penetration of AI into our everyday lives. - Different authors propose different recipes to achieve true intelligence, e.g., the author of this paper multiple computing and multiple knowledge [2]. In 2019, the IJCAI community proposed to merge deep ML technologies such as DNNs with models -Knowledge Representation (see Figure 7). It is therefore not primarily superintelligence or general intelligence, but the combination of DNNs with model-based (or rule-based) reasoning, that regardless of the upcoming problems at least in the near future will remain one of the dominant technologies and will persist as one of the most important technologies for the progress of human civilization. Believe it or not, even Google's responses to inquiries in recent years have been based on various AI methods. If AI stagnation should occur, then in the form of slower progress is expected, not of the winter type. - On the other hand, there is no doubt that in a few years’ time there will be another financial crisis, because on average crises occur in seven years, and in the last nine years we have been living in a series of steady growth. For the rich and clever, the crisis is an opportunity to enrich themselves more quickly, as history shows, but for the rest of the population, especially the middle and lower classes, there is nothing to cheer about. When and how deep the financial crisis will be -only time will tell. For the finale of the IJCAI 2019 report and for the overall impression, the last sentence of the Shoham’s lecture (Figure 8) was chosen: "This makes it the most exciting time to be an AI researcher." References [1] IJCAI 2019, https://ijcai19.org/ [2] Gams, M. 2001. Weak intelligence: through the principle and paradox of multiple knowledge. Nova Science. [3] Bostrom, N. 2014. Superintelligence – Paths, Dangers, Strategies. Oxford University Press, Oxford, UK. [4] Yampolskiy, R.V. 2016. Artificial Superintelligence. CRC Press. [5] Asilomar principles. 2017, (https://futureoflife.org/2017/01/17/principled-ai­ discussion-asilomar/). Consistency in Cloud-based Database Systems Zohra Mahfoud USTHB University, Algeria E-mail: mahfoud.zohra@yahoo.fr Nadia Nouali-Taboudjemat CERIST Research Center, Algeria E-mail: nnouali@cerist.dz Keywords: cloud computing, consistency, distributed databases, relational databases, No-SQL, CAP Received: July 15, 2019 Cloud computing covers the large spectrum of services available on the internet. Cloud services use replication to ensure high availability. Within database replication, various copies of the same data item are stored in different sites, this situation requires managing the consistency of the multiple copies. In fact, the requirement for consistency level can be different according to application natures and other metrics; a delay of some minutes in visualizing latest posts in social networks can be tolerated, while some seconds can make a loss of a bid in an auction system. Wide variety of database management systems are used actually by cloud services, they support different levels of consistency to meet the diversity of needs. This paper draws a presentation of the main characteristics of cloud computing and data management systems and describes different consistency models. Then it discusses the most famous cloud-based database management systems from the point of view of their data and consistency models. Povzetek: Prispevek analizira podatkovna skladišča v oblakih predvsem s stališča konsistentnosti. 1 Introduction Cloud computing refers to the large spectrum of services cloud systems and describes the implemented models of available on the internet. These services manage big data and consistency. Section 6 concludes the paper. quantities of data with high availability, scalability and elasticity. Providing availability requires databases 2 Cloud computing replication. Replication permits the creation and the Cloud computing includes all forms of services available management of various copies of data items stored in on the Internet; that are classified as software, platform different sites. or infrastructure as a service. Cloud services attract Consistency concerns the freshness of data and increasingly individuals, startups and big companies by indicates if copies are the same in the different sites and the fascinating characteristics offered such as witch version of data is returned by queries. In fact, Availability, Scalability and Elasticity [1-4]. consistency does not have the same importance for all the applications and the users. In social networks, a delay of minutes or even hours in visualizing the posts may not be 2.1 Availability a problem. Whilst for an auction system, a delay of few Queries must be answered within a reasonable time even seconds can cause the loss of a bid. there is a huge load of work or under any type of failures. Various systems are proposed to manage data for Availability is guaranteed by replicating databases, cloud services; they provide a variety of consistency i.e. creating multiple replicas (copies) of the database and models and use different data models which are based storing them at different sites. Replication can be full either on the classical relational model or on No-SQL when it concerns the entire database or partial when it models. concerns just a part of the database (one or more tables, This paper discusses consistency in cloud-based one or more partition) [5, 6]. database management systems. The reminder of this Typically, Replicas are used to increase the paper is organized as follows. Section 2 presents the availability of the system. They permit i) to decrease the main characteristics of cloud computing. Section 3 latency by distributing queries on different replicas, ii) to presents databases models in cloud. Section 4 explains cache site failure by accessing other sites, and iii) to the concept of consistency and the dilemma posed by the recover site failure as backups [7]. CAP theorem; it presents also the different levels and Synchronous replication control algorithms assume models of consistency. Section 5 presents some famous that replicas are the same all the time. But this is not possible physically, so outdated replicas are made not accessible until they are synchronized. In contrary, asynchronous algorithms allow accessing to divergent replicas that will finally converge [8, 9]. 2.2 Scalability This property is related to the capacity of providing large databases and managing their growing. Scalability is ensured by partitioning the database, i.e. devising the database into several disjoint partitions (fragments) that can be stored in different sites. Partitioning database offers the possibility of incrementing infinitely the capacity of storage by adding new hardware [6]. Partitioning has two general types: Vertical and Horizontal. In vertical type each partition contains a set of columns of the database; while in horizontal type (called communally sharding) the database is divided into sets of rows. The two types of partitioning can be combined to obtain a better strategy [10]. 2.3 Elasticity Elasticity called also elastic scalability refers to the flexibility of scaling up and down quickly in order to support the change of the requirements. Elasticity is the most important property that attracts companies to the cloud as it permits to pay accurately according to use. 3 Database models in cloud computing Data storage in the cloud uses both of the classical relational model and the new No-SQL architectures. Relational Databases: These databases respect the classical relational model proposed by E.F.Codd [11]. Relational databases structure data into tables composed of columns and rows, with a unique primary key and possible foreign keys. They provide the CRUD (Create, Read, Update and Delete) basic operations, and also operations across several tables. Relational databases dominate the market of databases for more than twenty years; this success is due to its stability and consistency. These characteristics are guaranteed via transactional mechanisms that are implemented by the ACID (Atomicity, Consistency, Isolation and Durability) properties [12]. SQL (Structured Query Language) is the most used for requesting and maintaining relational databases. Database Management Systems (DBMS) are responsible to store, retrieve, secure, replicate and realize backups of databases. The most famous Relational Databases Management Systems (RDBMS) are: Oracle, MySQL, Microsoft SQL Server, Postgres. No-SQL Databases: No-SQL databases (‘Not only SQL’ or ‘Not relational’) is a family of databases or more appropriate data stores that support all schemas of data characterized as structured, semi-structured and unstructured. No-SQL databases provide a high level of availability, scalability and elasticity. These features make No-SQL databases Z. Mahfoud et al. increasingly used for big data and qualified as the databases for the next-generation of web applications [13, 14]. Unlike the relational databases, No-SQL databases do not have a unified data model. Also, the level of operations is different; some systems provide only simple read-write operations, while others support more advanced operations. These differences lead to more than one hundred No-SQL databases which are principally classified into four categories [2, 3, 15, 23]: i. Key-value Databases: this model permits to store all schemas of data, as (key, value) pairs. A unique key is assigned to every value and permits to access the value. The value can be a simple data item, or a set of key-value pairs. Example of key-value databases are: App Engine Data Store, Redis, Riak, etc. ii. Column-oriented Databases: this model holds structured data in tables that are organized in rows like in relational databases. The difference is that columns can be different from one row to another. Also, a column can also regroup a set of columns. In other hand, operations across tables are not supported. Examples of column-oriented databases are: Google BigTable, Cassandra, etc. iii. Document-based Databases: this model is used to store unstructured data, where keys addressed generally XML (eXtensible Markup Language) or JSON (JavaScript Object Notation) documents. No restrictions on data type or documents length are imposed. Examples of Document-based stores are: CoucheDB, MongoDB, RavenDB, etc. iv. Graph Databases: this model allows storing data and relationships between them using graphs; nodes store data and arcs store relationships. The support of dynamic relationship makes this model the most appropriate for social networks. Examples of graph databases are: Neo4j, HyperGraphDB, Infinite Graph, etc. We stress that all No-SQL architectures are basically based on the key-value model. 4 Consistency Mutual consistency or simply consistency refers how to propagate updates between the different copies of replicated items. It concerns the state of data items in different sites; if they are the same or not. Also, how users see data items, if they see the same value, or they are allowed to see different values [6, 22]. Figure 01 shows a cloud system where the item X is duplicated in three sites. In an ideal situation, all copies of X have the same value (V1=V2=V3), this classical level of consistency is the most suitable, but it is hard to implement in distributed systems as it is proved by the CAP theorem as explained below. Figure 1: Distributed system with replication. We mention here that the cloud is considered as a large geo-distributed system; data is largely replicated to ensure availability in the case of concurrent queries and recovery in case of failure. The different replicas can be located in the same datacenter or over different geo-distributed datacenters that can located in different continents; in this case the communication between replicas is very expensive. 4.1 CAP theorem The CAP theorem (Figure 02) states that shared-data systems can ensure at most two of three properties: Consistency, Availability, and Partition tolerance at the same time [16, 17]. Choosing two properties between Availability, Partitioning tolerance and Consistency in the cloud is not easy; Availability and Partitioning are primordial and Consistency is vital for reliability. Cloud systems do not avoid absolutely one of the three properties, and propose generally a compromise between the three properties, which leads to support degraded levels of each one. A description of consistency levels is presented by the next section. PACELC [18] extends CAP and states that the compromise is not all the time between Availability and Partitioning and Consistency; during network Partition (P) the compromise is between Availability (A) and Consistency (C). Else (E), the compromise is between Latency (L) and Consistency (C). The latency measures the delay of getting a reply. 4.2 Consistency levels Consistency levels are influenced by the type of replication control protocol; i) Synchronous protocols propagate updates to all the replicas at the same time and Informatica 43 (2019) 313–319 315 in the same order. These protocols present strong consistency (immediate consistency). ii) Asynchronous protocols allow updating one replica while other outdated replicas are still accessible. iii) Hybrid protocols propagate updates synchronously between some replicas. Asynchronous and hybrid protocols present different levels of consistency according to which replicas are accessible, and the number of replicas that must be written and read before replying to queries [18, 19]. Quorum-based systems are proposed to achieve strong consistency by using the majority of replicas; Paxos is the most known protocol in this area [47]. The level of consistency is chosen according to the system nature and user’s needs. Transactional systems like they proposed to book a flight ticket, buy an item, or send a bid are cases where data must be treated with strong consistency; an inconsistency of few seconds may make a loss. Social networks are examples of applications that tolerate weak consistency; a delay in visualizing the latest posts can be accepted. 4.3 Consistency models A variety of consistency models degraded from strong to weak consistency are proposed in the literature, the main models are [19, 20, 21, 22, 40]: Strict consistency (Atomic consistency, Linearizability), is the strictest model of consistency; updates are propagated between replicas at the same order according to the real time. Also, reads return the last written values. Sequential consistency (Serializability): updates are ordered according to a logical order applied by all the replicas, this order can be different from the real order. Reads return the last values written according to the logical order. The eventual consistency model ensures that all replicas will eventually become consistent even if requests can read inconsistent values. Different variants of this model are distinguished according to the techniques used to manage the inconsistent window: Causal Consistency is a variation of the eventual consistency, where only causally related operations are ordered. Read-your-writes consistency is a case of causal consistency where users access always his updates, or a newer version, and never access an older version. Session consistency implements the read-your-writes consistency model during the session. The bounded staleness consistency model tolerates reading stale values under some conditions such as bounding staleness by a specific period of time delta. This condition is satisfied by propagating updates within delta. In Configurable consistency (Tunable consistency) the user configures the number of replicas accessed synchronously. Here, the consistency level depends on the percentage of the replicas requested synchronously; strong consistency is reached if the number of replicas for read (R) and write (W) overlap (R+W>=N), N is the total number of replicas. Consistency levels in cloud systems Wide variety of database management systems are used actually by cloud services. This section presents the most famous of them from the point of view of their data and consistency models [24]. 5.1 Amazon propositions Amazon has several propositions: Simple Storage Service (S3) [25,29], SimpleDB [26] and DynamoDB [27, 28] are No-SQL databases that provide high availability and scalability. Amazon Aurora [30, 31] is a relational databases management system that provides strong consistency. S3 is designed to store large data in buckets: a bucket is organized as a key-value store, values are generally objects that represent data files or folders used to organize data files, folders can be arranged hierarchically. S3 offers simple operations to create, write, read and delete buckets, keys and objects. S3 uses automatic Cross-region replication that allows asynchronous copying of objects across buckets in different Regions. This strategy provides eventual consistency model. SimpleDB arranges structured data in domains which consist of items; items are composed of pairs of (attribute, value); value can contain multiple data. SimpleDB offers operations for creating, writing, reading and deleting a domain or an attribute. Operations manipulate one or various items of the same domain. Eventual consistency is proposed by default; however it is possible to choose the strong consistency. Dynamo uses tables of items, each item contains one or more attributes. An attribute is composed of (key, value) pairs. Dynamo provides several operations to create, write, read and dele table, item and attribute; which permit to manipulate one or various items of the same table. Initially, dynamo offers eventual consistency; a quorum that preserves availability and scalability is addressed to fulfill operations. However, dynamo makes it is possible to achieve strong consistency by configuring the number of requested replicas. Data models in simpleDB and Dynamo have the structure of tables. Although, they are not classified as column-family store because they have simple columns and not super column families. Amazon Aurora is a cloud-based relational databases management system proposed by Amazon Relational Database Service (RDS). Aurora is built on a MySQL engine and it is compatible with PostgreSQL. It provides better availability and scalability comparing to classical databases engines on RDS. Aurora guarantees strong consistency by supporting a quorum protocol. 5.2 Google propositions In its turn, Google published several cloud-based systems [32] like BigTable [33], Megastore [34], Spanner [35], Cloud SQL [36], and Cloud datastore [37]. Bigtable stores data in massive tables. Each table is organized in rows that are accessed by primary keys and Z. Mahfoud et al. they contain a set of column-families which can differ from a row to another. A column-family regroups related columns and each column contains a single value for a row. This model allows storing versioned data in columns regrouped in a column family. Operations concern atomic single-row and a quorum protocol based on Paxos algorithm is implemented to provide strong consistency for write operations, read operations can get stale data if an update is on progress. Bigtable is designed to store very large amounts of data; Google uses it in many applications like: Google Analytics, Earth, Map and Personalized Search. Megastore uses schemas of tables to organize data; a table contains a set of entities that are characterized by a set of properties. Megastore defines entity groups that are sets of related tables based on Bigtable. Megastore provides transactions with full ACID semantics that can concern data through several tables of the same entity group, not just data of the same table like the majority of No-SQL databases. Like Bigtable, Megastore uses Paxos protocol to provide strong consistency; for each write operation, a majority of replicas across geographically distributed datacenters is requested; this strategy increases the system latency. Megastore is proposed to build interactive applications; it is used by well-known Google applications as: AppEngine, Gmail, Calendar and Android Market. Spanner is a key-value database created to fix the weaknesses of megastore in term of latency. Like megastore, spanner organizes data in schematized semi-relational tables, uses timestamp for versioning data and use a like SQL-based query language. Spanner propose an excellent support of transactions with full ACID properties, it provide strong consistency for distributed transactions across geographically replicated datacenters; this is achieved by executing a combination of the two-phase-commit protocol and Paxos protocol. Spanner is largely used within Google's datacenters infrastructures. Cloud SQL is a RDBMS based on MySQL that provides classically immediate consistency. Cloud Datastore is a Document store that organizes data on kinds of entities; each entity is accessed by a key and composed of a set of properties storing values that can have different types even for the same properties. Cloud Datastore use Multi-Master replication based on Paxos. Queries are configured to obtain immediate or eventual consistency. 5.3 Microsoft propositions Microsoft has also several propositions: Microsoft Azure Table storage [38], Microsoft Azure DocumentDB [39] and Microsoft Azure SQL Database [41]. Microsoft Azure Table storage is a key-value store that stocks large amounts of data in tables. Each table contains a set of entities: an entity is composed of a primary key and a set of properties. Table storage provides strong consistency, and permits to achieve transactions with ACID properties across tables of the same partition. Microsoft Azure Cosmos DB gathers multiple data models that include key-value, table, columnar, document and graph data models. It offers a configurable consistency model that presents five levels: strong, bounded-staleness, session, consistent prefix, and eventual. Strong consistency is associated only with one Azure region; it uses a linearizability based on a majority of replicas. The other levels are designed to reinforce avalability across different regions. Microsoft Azure SQL Database is a RDBMS in the cloud built on the Microsoft SQL Server engine that supports full ACID properties of relational databases and uses a quorum-based algorithm that provides an acceptable consistency level with high availability. 5.4 Others solutions 5.4.1 Cassandra Cassandra [42, 43] is an open source column family store proposed by Facebook for managing massive amounts of data. Cassandra is inspired from Google BigTable and Amazon DynamoDB. The data model of Cassandra uses column families (tables) that regroup rows; each row in a table is composed of a key and a list of columns or super columns. A column is composed of a key, a value and a timestamp. A super column is a column family that regroups columns. Cassandra proposes panoply of consistency models that can be configured at operation level. These levels are differentiated according to the requested replicas and theirs locations; the level ALL involves all the replicas of the cluster. The levels: One, TWO and THREE involve at least one, two and three replica (s), respectively. The level QUORUM involves a quorum of replicas of the cluster. According to the nodes locations, the following levels are defined: EACH_QUORUM requires a quorum of replicas in all data centers. LOCAL_QUORUM requires a quorum of replicas in the same data center. And, LOCAL_ONE requires one replica at least in the local data center. In addition, Cassandra proposes the SERIAL level that uses linearizable consistency for achieving lightweight transactions. LOCAL_SERIAL concerns one datacenter. The levels listed above are common to read and write operations. The ANY level is specific only to write operations; it permits to execute a write operation even if no required replica is available; the operation writes hints for downed nodes on others nodes. The changes will be sent to downed nodes when they recovered. The consistency level is determined by the number of replicas solicited for the read (R) and write (W) operations; if it overlaps the total number of replicas (N) the consistency is strong (R+W>=N), otherwise the consistency is weak. 5.4.2 PNUTS PNUTS [44, 45] proposed by Yahoo! exposes a simple relational model with flexible schema. PNUTS organizes Informatica 43 (2019) 313–319 317 data into tables of records with attributes that can store any type of data. PNUTS offers various operations like Update, delete, selection of one or more items from a single table. PNUTS proposes a per-record timeline consistency model that offers a consistent view of data to the user; a master replica is nominated to each record, this replica receives all the updates concerning the record and propagates the updates to other replicas in the same order. This consistency can be configured; the weak level is ensured by the options: Read-any, Read-critical (required version), Test-and-set-write. However, the options: Read-latest ensures strong consistency. 5.4.3 Neo4j Neo4j [46] is a graph based No-SQL databases that models data using nodes and relationships. Nodes are used to represent entities, they can be labeled and contain properties. Relationships present relations between nodes and can also contain properties. Neo4j supports full ACID properties and implements causal consistency to provide an acceptable level of consistency. 6 Conclusion Availability, scalability and elasticity are the success keys of cloud computing. At the storage level, these properties are guaranteed by partitioning and replicating databases. Initially, cloud systems used the relational model that dominated the market of databases for more than twenty years. This model is known by its stability and consistency, which are guaranteed using transactional mechanisms. However, these mechanisms make the relational model very rigid and lack required availability and scalability. In order to meet the cloud needs, a new generation of relational cloud-based systems that supports more availability and scalability appeared. Several applications in cloud prefer No-SQL models that are proposed initially as simple key-value pairs that avoid all types of constraints. Bit by bit, No-SQL Databases use more organized models and integrate some transactional mechanism. Nevertheless, they still more flexible comparing to relational model. In the consistency side and as it is difficult to ensure availability with strong consistency in large geo-distributed systems, cloud systems implement different consistency models to ensure the best compromise between availability and consistency. In addition, a lot of systems propose a tunable consistency that offers the possibility to choose between numerous proposed models. 7 References [1] S. Sakr, A. Liu, D. Batista, M. Alomari (2011). “A Survey of Large Scale Data Management Approaches in Cloud Environments”. IEEE Communications Surveys and Tutorials. 13(3): 311-336, https://doi.org/10.1109/SURV.2011.032211.00087. [2] A. Elzeiny, A. Abo Elfetouh ,and A Riad (2013). “Cloud Storage: A Survey”. International Journal of Emerging Trends & Technology in Computer Science. Vol. 2, Issue 4, ISSN 2278-6856: 342­349. [3] M. Siba, S. Breß, and E. Schallehn (2012). "Cloud Data Management: A Short Overview and Comparison of Current Approaches". Grundlagen von Datenbanken. [4] D. Kossmann, T. Kraska, S. Loesing (2010). “An evaluation of alternative architectures for transaction processing in the cloud”. SIGMOD Conference : 579-590. https://doi.org/10.1145/1807167.1807231. [5] Saeed K. Rahimi , By (author) Frank S. Haug (2010). “Distributed Database Management Systems A Practical Approach”. Wiley-IEEE Computer Society. https://doi.org/10.1002/9780470602379. [6] M.T Özsu, P. Valduriez (2011). “Principles of Distributed Database Systems”. Springer Science+ Business Media, 3rd ed. https://doi.org/10.1007/978-1-4419-8834-8. [7] V.K. Pallaw (2010). “Concept of Database Management Systems”. Asian Books Pvt. Ltd. ISBN : 978-81-8412-119-3. [8] M. Wiesmann, F. Pedone, A. Schiper, B. Kemme, G. Alonso (2000). “Understanding Replication in Databases and Distributed Systems”. IEEE International Conference on Distributed Computing Systems: 464-474. [9] M. Wiesmann, F. Pedone, A. Schiper (2000). “Database Replication Techniques: a Three Parameter Classification”. The IEEE 19th Symposium on Reliable Distributed Systems: 206­215. [10] SH. Navathe, S. Ceri, G. Wiederhold, J. Dou (1984). “Vertical Partitioning Algorithms for Database Design”. ACM Transactions on Database Systems, Vol. 9, No.4. https://doi.org/10.1145/1994.2209. [11] Codd, E.F. (1970). "A Relational Model of Data for Large Shared Data Banks". Communications of the ACM. 13 (6): 377–387. https://doi.org/10.1145/362384.362685. [12] J. Gray (1981). “The Transaction Concept: Virtues and Limitations”. The 7th VLDB, Cannes: 144-154. [13] F. Bugiotti, L. Cabibbo, P. Atzeni, R. Torlone (2014). “Database Design for NoSQL Systems”. 223-231. [14] P. J. Sadalage and M. J. Fowler (2012). “NoSQL Distilled”. Addison-Wesley. [15] G. Harrison (2015). “Next Generation Databases: NoSQL, NewSQL, and Big Data”. Apress, ISBN(e): 978-1-4842-1329-2. [16] E. A. Brewer (2000). “Towards Robust Distributed Systems”. PODC (Invited Talk) :7. Z. Mahfoud et al. [17] N. Lynch and S. Gilbert (2002). “Brewer's conjecture and the feasibility of consistent, available, partition-tolerant web services”. ACM SIGACT News, Vol. 33 Issue 2: 51-59. https://doi.org/10.1145/564585.564601. [18] Daniel J. Abadi (2012). “Consistency tradeoffs in modern distributed database system design: Cap is only part of the story”. Journal of computer, 45(2):37–42. https://doi.org/10.1109/MC.2012.33. [19] S.P. Kumar (2016). “Adaptive Consistency Protocols for Replicated Data in Modern Storage Systems with a High Degree of Elasticity”. PHD thesis, Conservatoire national des arts et métiers, Paris, France. [20] D. Mosberger (1993). “Memory Consistency Models”. ACM SIGOPS Operating Systems Review Homepage archive. Vol. 27, Issue 1 : 18-26 https://doi.org/10.1145/160551.160553. [21] Adve, Sarita V and Gharachorloo, Kourosh (1996). “Shared Memory Consistency Models: A Tutorial”. Journal of Computer, Vol. 29, Issue 12: 66-76. https://doi.org/10.1109/2.546611 [22] W. Vogels (2009). “Eventually consistent”. Communications of the ACM, Vol. 52, n.1: 40-44, https://doi.org/10.1145/1435417.1435432. [23] IGI Global publications (2016). “Big Data: Concepts, Methodologies, Tools, and Applications”. ISBN: 9781466698406. [24] “DB-Engines Ranking”, Available Online [Aug2018]: http://db-engines.com/en/ranking/. [25] “Amazon Simple Storage Service Documentation”. Available Online [Aug2018]: https://aws.amazon.com/documentation/s3/. [26] “Amazon SimpleDB Documentation”. Available Online [Aug2018]: https://aws.amazon.com/documentation/simpledb/. [27] “Amazon DynamoDB Documentation”. Available Online [Aug2018]: https://aws.amazon.com/documentation/dynamodb/. [28] G. DeCandia, D. Hastorun, M. Jampani, et al. (2007). “Dynamo: Amazon’s highly available key-value store”. SOSP:205–220. https://doi.org/10.1145/1294261.1294281. [29] D. Bermbach and S. Tai (2011). “Eventual consistency: How soon is eventual? an evaluation of amazon s3’s consistency behavior”. The 6th Workshop on Middleware for Service Oriented Computing. ACM. https://doi.org/10.1145/2093185.2093186. [30] “Amazon Amazon Aurora”. Available Online [Aug2018]: https://aws.amazon.com/rds/aurora/. [31] “Amazon Relational Database Service Documentation”. Available Online [Feb2017]: https://aws.amazon.com/documentation/rds/. [32] “Google Cloud Platform: Cloud Storage Products”. Available Online [Aug2018]: https://cloud.google.com/products/storage. [33] F. Chang, J. Dean, S. Ghemawat, et al. (2008). “Bigtable: A Distributed Storage System for Structured Data”. ACM TOCS 26.2, 4:1–4:26. https://doi.org/10.1145/1365815.1365816. [34] J. Baker, C. Bond, J. Corbett et al. (2011). “Megastore: Providing Scalable, Highly Available Storage for Interactive Services”. CIDR: 223–234. [35] J. Corbett, J. Dean, M. Epstein, et al. (2012). “Spanner: Google’s globally-distributed database”. OSDI:251–264. DOI: 10.1145/2491245. [36] “CLOUD SQL”. Available Online [Feb2017]: https://cloud.google.com/sql/. [37] “Google Cloud Datastore Documentation”. Available Online [Aug2018]: https://cloud.google.com/datastore/docs/. [38] B. Calder, J. Wang, A. Ogus et al. (2011). “Windows Azure Storage: A Highly Available Cloud Storage Service with Strong Consistency”. The 23rd ACM Symposium on Operating Systems Principles: 23-26. Cascais, Portugal. [39] “Azure Cosmos DB Documentation”. Available Online [Aug2018]: https://docs.microsoft.com/en-us/azure/cosmos-db/. [40] A. Singla, U. Ramachandran, and J. Hodgins (1997). “Temporal Notions of Synchronization and Consistency in Beehive”. The 9th Annual ACM Symp. on Parallel Algorithms and Architectures: 211–220. https://doi.org/10.1145/258492.258513. [41] “Microsoft Azure SQL Database”. Available Online [Aug2018]: https://azure.microsoft.com/en­ us/services/sql-database/ [42] “Apache Cassandra”. Available Online [Aug2018]: http://cassandra.apache.org/ [43] A. Lakshman, P. Malik (2010). “Cassandra: a decentralized structured storage system”. Operating Systems Review 44(2): 35-40. https://doi.org/10.1145/1773912.1773922. [44] B. Cooper, R. Ramakrishnan, U. Srivastava (2008). “Pnuts: Yahoo!’s hosted data serving platform”. PVLDB, 1(2):1277–1288. https://doi.org/10.14778/1454159.1454167. [45] A. Silberstein, J. Chen, D. Lomax et al. (2012). “PNUTS in Flight: Web-Scale Data Serving at Yahoo”. IEEE Internet Computing 16(1): 13-23 https://doi.org/10.1109/MIC.2011.142. [46] “Neo4j”. Available Online [Aug2018]: https://neo4j.com/ [47] L. Lamport (2002). “Paxos Made Simple, Fast, and Byzantine”. OPODIS: 7-9 Output Analysis in Voice Interaction in AI Environment Fanyu Jin Yancheng Institute of Technology, Yancheng224051, Jiangsu Province, China E-mail: tomjin2001@126.com Keywords: AI interface; voice interaction; output analysis; cultural elements Received: July 15, 2019 The future foreign language teaching will inevitably be combined with AI technology, and it is likely that the traditional foreign language teaching method of one teacher instructing a number of students will gradually be completely replaced by a new foreign language learning mode of each student’s foreign language learning and most importantly acquisition being realized by interacting with AI interface customized for each student. The reason is rather simple. In spite of numerous repatching teaching method explorations, the traditional classroom-based foreign language teaching has been unable to solve the congenital problems, such as inadequacy of language input and interaction, insufficiency of real life sensory stimulation and violation of natural language acquisition sequence for the absence of language environment. The AI interface, in contrast, with the infinite, accurate and real language supply and human-computer interaction, and also with constant adjustment of ZPD (Zone of Proximal Development) according to each student's language development level, precisely sets up appropriate scaffolding for every language learner, thus revolutionarily creating a language environment close to or even beyond the real one and returning language learning to natural acquisition process. The first step to achieve this goal is to realize human-computer voice interaction. The realization of voice interaction needs many technical supports, among which voice interaction output analysis is an urgent part. By importing AI voice interactive output analysis algorithm, constructing output analysis model, and establishing the operation platform of the analysis model, the paper relies on the determination of the voice interactive output influence function, and takes the cultural elements of English language as an example to analyze the output. Povzetek: V članku je objavljana analiza govorne komunikacije z upoštevanjem mehanizmov angleškega kulturnega okolja. 1 Introduction Interaction based on mechanical one-way input have analysis of voice interaction in AI environment. In order been giving way to the two-way voice interaction in to ensure the validity of the interactive output analysis numerous ways to satisfy human-computer method and simulate the language interaction communication needs especially in language learning. environment in AI environment, two different methods For instance, on the basis of Google search function, of interactive output analysis are used to analyze the Google Now records the keywords searched by users, output accuracy simulation experiments. It is found that and provides users with relevant voice services through the method in this paper has higher analysis output intelligent reading. This allows the machine to upgrade accuracy [1]. from “passive” answering user's questions to “active” alerting users to their needs, that is, the way the machine 2 Construction of AI voice interacts with human beings in a service-oriented interactive output analysis model manner. Whether it’s Apple AI or Google Now, it gives machines the ability to act on the basis of “independent At present, the internationally recognized and accepted thinking”, thus opening a new era of language learning communicative teaching method aims to cultivate with two-way human-computer interaction. Conventional students' ability to express themselves in the target voice interactive output analysis method uses dynamic language. Interactive teaching method is a kind of voice capture technology to realize voice interactive communicative language teaching. Interactive teaching output analysis, which can greatly improve the efficiency theory holds that language is the system of expressing of voice output. However, when applied to voice ideological system, and the main function of language is interactive output analysis in Artificial Intelligence (AI) interaction and communication. Interaction mainly refers environment, due to the high degree of strangeness in the to the interaction between teachers and students and also field and the limited response of the operating among students themselves in the classroom. Classroom environment, the problem of low accuracy of output activities should include real communication and enable analysis appears. This greatly reduces the frequency of students to perform meaningful tasks. Students' AI environment use, resulting in incomplete output communication includes sharing information and negotiating meanings with others. The construction of AI voice interactive output analysis model mainly includes two parts: building the running platform of AI voice interactive output analysis model and importing AI voice interactive output analysis algorithm. 2.1 Establishment of AI voice interactive output analysis model platform The model of Multi-dimensional Input-interactive Output is based on the premise of “language learning is technical training, not pure knowledge learning”, and guided by Robert W. Blair's Low Shielding Effect in 1987 and Swain's Output Hypothesis in 1995. In the process of building this model, we will focus on the real language input mode and the learner's output mode which play decisive roles in the latter language output. The running platform of AI voice interactive output analysis model is the basic platform to ensure the reasonable and safe operation of AI voice interactive output analysis model. The platform consists of four parts: data layer, operation layer, physical layer and display layer [2]. Data layer is the logical level for acquiring voice interactive information, which provides data support and data guarantee for operation layer. Under the premise of authorization, the security environment is guaranteed, the voice interaction status and voice interaction information are acquired, and the data supply is completed. Operational layer is a logical level based on AI voice interaction output analysis algorithm, which provides direct and indirect evidence for output analysis in AI voice interaction. The physical layer includes all external devices supporting the platform, such as processors, hosts, displays, network connections, etc. It provides a hardware platform for the platform and output analysis, in AI environment. The display layer is the logical level through which the results are displayed by the display devices in the physical layer, so that the staff can read the results directly and realize the output analysis in the voice interaction in AI environment [3]. 2.2 An interactive output analysis algorithm for importing AI speech For the studying the output analysis of voice interaction in AI environment, the algorithm of AI voice interaction output analysis’ main function is to analyze the speech interaction output of artificial intelligence. In the environment of voice interaction, taking the cultural elements of English language as an example, AI voice interactive output analysis algorithm can not only drive mathematical operations, but also complete the transforming voice into written language. In the process of data calculation, the conversion process from voice to language and the process of driving mathematical operation are analyzed in detail. F. Jin In the process of voice-to-language transformation, the parameters are generated and the results are output mainly through voice acquisition module and voice recognition module. With different results of the output, the command is written in the form of MySQL statement. The key program of the command is shown in Figure 1. Figure 1: Key program of AI voice interactive output analysis algorithm. 3 Output analysis in implementing voice interaction in AI environment Based on the construction of AI voice interactive output analysis model, the influence function of voice interactive output is determined. Taking the cultural elements of English language as an example, the output analysis is carried out to realize the output analysis of voice interaction in AI environment. To determine the influence function of voice interactive output, we need to determine the state equation of voice interaction and calculate the interaction coefficient of voice, so as to realize the determination of the influence function of voice interactive output. The speech interaction state is real-time interaction, and is difficult to quantify, thus it is necessary to construct the speech interaction state equation. Different types of language interaction have different operational systems of equations and methods of constructing equations. This paper takes the voice interaction between English users as an example to construct the state equation of voice interaction. Based on the relevant information acquisition module of AI voice interactive output analysis model operation platform, the AI voice interactive output analysis algorithm is imported for statistical analysis. Assuming that the amount of interactive information is M and the interaction coefficient q is 1.0, there is a state equation of inter-cell voice interaction, as is shown in formula (1)[4]. ( 1) In the formula, C represents the state equation of inter-cell voice interaction, and qx represents the type of language, and T represents the time of voice output. When the interaction mode is more complex, the interaction coefficient q can be expressed by formula (2): (2) In the formula, Q represents the ideal coefficient, C represents the state equation of inter-cell voice interaction, and .T represents the type of interaction mode. Through the determination of the speech interaction state equation, the speech interaction coefficient is calculated relying on the variable relationship of the speech interaction state equation. The calculation of speech interaction coefficient is based on the interaction equation. The calculation process is as follows[5]: (3) In the formula, D represents the interactive state, V represents the interactive output, and F represents the scope of language interaction[6]. When the speech interaction coefficient does not satisfy the interactive equation, the statistical calculation of the data is carried out, as is shown in formula (4). (4) The condition that the speech interaction coefficient does not satisfy the interaction equation is as follows: Speech characters are less than 4 characters, there are uncommon words in the speech which cannot be recognized, and the speech discrimination is not high, making it difficult to recognize[7]. Based on the calculation of voice interaction coefficient, the influence function of voice interaction output is determined, which can be expressed by formula (5): (5) In the formula, g is an interactive way, such as English. f is running, and k represents the reliability of interactive running program[8]. The determination of the influence function of the speech interactive output analysis algorithm is realized by the formulas (1)~(5). Example analysis In order to ensure the output analysis in the speech interaction in AI environment proposed in this paper, the cultural elements in the English language are used as examples, and the English audio interaction in different AI environments is used as the test object to analyze the output precision simulation experiment[9]. Different cultural forms of English sound interaction in AI environment, speech types, dialects, etc. are simulated [10]. The simulation experiment was carried out by using the conventional interactive output analysis method as the experimental comparison object[11]. 4.1 Test data preparation In order to ensure the accuracy of the analysis output simulation test, the experimental data preparation is first carried out, and the SDH-214 simulation system is Informatica 43 (2019) 321–324 323 selected for the data operation platform [12]. On the two sets of identical computers, two different interactive output analysis methods were used to analyze the output precision simulation [13]. The experiment mainly consisted of English speech interaction, with different English customs and cultural characteristics for analysis[14]. 4.2 Test results analysis During the experiment, two different methods of interactive output analysis were used to analyze the change of output accuracy in the simulation environment. The simulation curves of analytical output accuracy are obtained, as is shown in Figure 2. According to the analysis of the test curve results, the output accuracy of the proposed interactive output analysis method is 73.45%, and that of the conventional interactive output analysis method is 39.24%. Compared with the conventional interactive output analysis method, the output accuracy of the proposed interactive output analysis method is 34.21%, which is suitable for the output analysis of voice interaction in AI environment [15]. 5 Conclusion This paper presents the output analysis of voice interaction in AI environment with examples of cultural elements in language. The research conducted on the base of the construction of AI voice interactive output analysis model and the determination of relevant parameters. The experimental data show that the proposed interactive output analysis method has high analysis output accuracy. This study provides a new idea for the interactive output analysis method, as well as a theoretical basis for the interactive output analysis method, and lays a foundation for the further study of voice interaction analysis. However, there are still some deficiencies in the practical application of this paper. The author hopes to further improve the output accuracy of voice interaction in the future research. References [1] KimHC . Weaknesses of Voice Interaction[C]// International Conference on Networked Computing & Advanced Information Management. IEEE, 2008. [2] KostovV , FukudaS . Emotion in user interface, voice interaction system[C]// IEEE International Conference on Systems. IEEE, 2000. [3] Shriver S ,TothA , ZhuX , et al. A Unified Design for Human-Machine Voice Interaction[C]// Chi 01 Extended Abstracts on Human Factors in Computing Systems. ACM, 2001. [4] Osawa H , Orszulak J , Godfrey K M , et al. Improving voice interaction for older people using an attachable gesture robot[C]// Ro-man. IEEE, 2010. [5] MassieT ,WijesekeraD . TVIS: Tactical Voice Interaction Services for dismounted urban operations[C]// Milcom IEEE Military Communications Conference. IEEE, 2013. [6] LeeA ,OuraK , TokudaK . Mmdagent—A fully open-source toolkit for voice interaction systems[C]// IEEE International Conference on Acoustics. IEEE, 2013. [7] DrigasA ,ArgyriK , VrettarosJ . Decade Review (1999-2009): Artificial Intelligence Techniques in Student Modeling.[C]// Best Practices for the Knowledge Society Knowledge, Learning, Development & Technology for All, Second World Summit on the Knowledge Society, Wsks, Chania, Crete, Greece, September. DBLP, 2009. [8] GorostizaJF ,SalichsMA . Teaching sequences to a social robot by voice interaction[C]// IEEE International Symposium on Robot & Human Interactive Communication. IEEE, 2009. [9] WolfgangGarn,JamesAitken. Agile factorial production for a single manufacturing line with multiple products[J]. European Journal of Operational Research,2015(3). [10] JamesAitken,CecilBozarth,WolfgangGarn. To eliminate or absorb supply chain complexity: a conceptual model and case study[J]. Supply Chain Management: An International Journal,2016(6). [11] AfshinSamani,DivyaSrinivasan,SvendErikMathiass en,PascalMadeleine. Nonlinear metrics assessing motor variability in a standardized pipetting task: Between-and within-subject variance components[J]. Journal of Electromyography and Kinesiology,2015(3). [12] Matsushita Y , Uchiya T , Nishimuray R , et al. Crowdsourcing Environment to Create Voice Interaction Scenario of Spoken Dialogue System[C]// International Conference on Network-based Information Systems. IEEE, 2015. [13] ZhouF Y ,Li JH,TianG H ,et al.Researchand Implementation of Embedded Voice Interaction System Based on ARM in Intelligent Space[J]. Advanced Materials Research, 2012, 433­440:5620-5627. F. Jin [14] MitreaD ,Mitrea M . Voice interaction in an integrated office and telecommunications environment[J]. J.am.chem.soc, 2015, 111(8):1138­1157(20). [15] Coppola P ,Mea V D , GasperoLD ,et al. AI Techniques in a Context-Aware Ubiquitous Environment[M]// Pervasive Computing,Springer London, 2009. Research on the Simulation Design of Humanistic Landscape Optimization in Urban Residential Area Based on Computer Technology Wei Wu1,2, Boxun Wang2 and Shuai Yang2 1He Xiangning College of Art and Design, Zhongkai University of Agriculture and Engineering Guangzhou 519000, Guangdong Province, China https://www.zhku.edu.cn 2Faculty of Innovation and Design, City University of Macau, Macau http://www.cityu.edu.mo E-mail: 50958250@qq.com Keywords: computer technology, urban residential area, humanistic landscape, a model of optimization design, future urban landscape design Received: July 15, 2019 With economic development, individuals are paying increasing attention to their surrounding living setting. They attach excellent significance to urban design coordination and the general natural environment, and follow a lifestyle that is easy and comfortable. In the past, scientists concentrated on the physical shape of residential areas and landscape design, and few quantitative research on ecological housing fields ' color modifications have been conducted. The human landscape optimization design of urban residential areas is performed on the basis of computer technology. The urban landscape design of residential urban landscapes was carried out from various stages. Comparison is made between the three-dimensional color and three-dimensional color rates of two distinct model plant color landscape models, and the fundamental requirements for rational landscape color plant distribution are acquired. A computer-based model of artificial landscape layout is suggested for urban residential areas. It offers a theoretical foundation for future design of urban landscape Povzetek: Predstavljena je raziskava urbanega okolja v smislu optimizacije več parametrov. Introduction The residential areas in almost every city will have more or less humanistic landscape. These humanistic landscapes are the organic combination of the humanities and the landscape. Humane landscape is one of the indispensable conditions to improve the quality of urban life. It enables residents to enjoy the natural scenery and the aesthetic pleasure, feel the edification of culture and improve the taste of life in daily life. For example, some greening and water bodies in the city, as well as various facilities, not only have their basic material functions, for example, green landscaping can improve urban greening, absorb carbon dioxide, clean air, and reduce soil erosion, but also has the function of spirit in the shape and design, they are more artistic, which can give people bring aesthetic pleasure enjoy. Through humanistic landscape design, the material and spiritual functions of these landscapes can be maximized. Because the humanities landscape design not only can show the integrity and ecology of the landscape, but also fully display its artistic and comfortable. The design of humanistic landscape in urban residential area accords with the needs of modern people's living environment, improves people's living environment, and is also conducive to the protection of natural environment, so that people can get along well with the environment [1]. 2 Literature review IMcharg has ever put forward an important idea of the design of urban residential environment and humanistic landscape, namely the idea of comprehensive ecological planning. In the book“water landscape”, Rosemarie Mike Lily focuses on the design of the garden waterscape. In the book “ecological design and comprehensive treatment of urban waterscape” NARS: creating a clear and beautiful original waterscape system, JinyuanHuan focuses on the analysis of waterscape's ecological design concept and how to carry out water ecological design. He pointed out the shortage of modern urban waterscape, analyzed the reasons for the problems, and put forward the “NARS ecological waterscape system”. BaiYufang and Chen Wangqing discussed the application of the virtual Waterscape in the application of the waterscape of Hangzhou residential district [2]. In recent years, great achievements have been made in the construction of urban residential areas in China. However, it is also faced with the problems of large residential area, excessive consumption of resources and energy, and low ecological degree of residential areas. Therefore, while China is actively promoting the “low-carbon economy” growth mode, increasing the proportion of ecotype residential areas is the inevitable direction of the development of residential areas. At present, the Figure 1: Principle of humanistic landscape design in urban residential area. research on Eco residential area in China is not deep enough. Under the urgent task of building eco residential area, we need to strengthen the research on the theory of ecological residential area. Figure 1 is the principle of humanistic landscape design in urban residential areas. 3 Research methodology Optimization simulation design of humanistic landscape in urban residential area based on computer technology 3.1 Design principle In the process of three-dimensional image design of humanistic landscape in urban residential area, the urban landscape and landforms are obtained first, and the characteristics of humanistic landscape plant configuration are obtained. The stereoscopic index of plant colorization in urban residential area is set up, and the three-dimensional image design of humanistic landscape in urban residential area is completed based on this. The detailed steps are as follows: CI represents the environmental factors of the humanistic landscape before the design. Hp represents the ecological factors before the design of city residential landscape. MF represents the natural factors. represents human factors. [3] Formula (1) is used to obtain the characteristics of humanistic landscape configuration. (....·....) .... . = .... (1) ..~×.. In the formula, represents the function of humanistic landscape in urban residential area, and represents the physiological and ecological characteristics of color plants. Assuming that ..( ..) represents the equal number of color plants, C(i,t) represents the physical form of urban road landscape.The formula (2) is used to give the three-dimensional index of the color of urban plants. Figure 2: The rational collocation of color plants in the cultural landscape. ..(..)×..(..,..) ....(..)= ...(....) (2) .....(..)·..(I) ..,.. In the formula, represents the physical effect of the color of urban landscape plants. represents the greening level of the city. is the characteristics of leaf color leafed plants. hj represents the color leafed plants leaves. represents the proportion of different colors of leaf plants to the whole. Mk represents the comprehensive evaluation of color leafed plants adaptability equation. The formula (3) is used to form a three-dimensional image design model of the urban landscape plant landscape. ..(....) ... (....)= ×[h..·..(....)].... (3) .... . ×....(..) In the formula, ..(....) represents the variety of color plants and their growth adaptability. However, traditional methods do not consider the seasonal variation of color and the physiological and ecological characteristics of plants. 3.2 Three-dimensional image optimization design of humanistic landscape in urban residential area (1) Rational collocation of color plants in the cultural landscape Plants commonly used in urban greening plants contain various kinds of trees, herbs and shrubs. They are the basic elements of urban ecological and cultural environment, as shown in Figure 2. According to the current urban environmental conditions and existing plant resources and their Greening Status, color symbolization, color Psychological Association, the configuration is carried out in the process of establishing the optimization model of the humanistic landscape design. In the configuration of color plants, seasonal variation of plant biology characteristics and color are combined. And according to the constraint conditions of humanistic landscape design in the residential area of urban residents, the landscape elements of the landscaping of the plants are obtained. The detailed steps are as follows: v(ol) represents the natural landscape elements of the city before the landscape design. represents the environmental conditions of the city. ku represents existing plant resources. represents the main body of the urban green space system. The formula (4) is used to obtain the basic conditions for the rational collocation of color plants in the landscape. ..(..)·..(....) ..(..) ..(..,..)= . ..(....) (4) ......×...... ..(..,..) In the formula, f(XC)represents the physical effect of plant color. HHk represents the symbolization of plant color. k(w) represents the geographical latitude and terrain of the greening site. represents the coordination among plant populations, and ..(....)represents the species of color plants. It is assumed that h(uy) represents the basic principles of rational collocation of plants. m(l) represents the seasonal variation rule of color plants. The formula (5) is used to get the elements of landscaping. ..(..)×h(....) ....(..,..)= (5) ..(..)·..(..) In the formula, ..(..) represents the mutual coordination between the color plant population, and ..(..) represents all the features of the landscape. It is assumed that .... is the form of plant color expression. kl represents the effect of color patches in urban humanistic landscape, and the color and vegetation characteristics of color plants are obtained by using formula 6. ..(..)×....*(..).... ..(....) ..(..)*= ·..(..,..). (6) ....×....(....).... ....(..,..) In the formula, ..(..)represents the urban green space system. jk represents the color leafed plants adaptability. is the city lottery leaf plant resources. v represents the principle of urban characteristics. bn(lp) represents plant planting requirements. The above analysis can show that in the process of establishing the optimization model of the urban humanistic landscape plant color design, the basic conditions for reasonable collocation of plants in the landscape are obtained by the principle of highlighting the urban characteristics and ecology. According to the needs of the landscape theme, the colorful landscape elements of the landscape plants are obtained, and the seasonal variation of color plants is obtained. The physiological and ecological characteristics of the color plants are given, which lays the foundation for the optimization design of the three-dimensional image of the urban landscape plant landscape [4]. (2) Optimization design of three-dimensional image of humanistic landscape in urban residential area Based on the physiological and ecological characteristics of color plants obtained above, the three-dimensional color quantity concept is proposed based on the three-dimensional image optimization design of plant landscape in urban cultural landscape, and the three-dimensional color amount of urban humanistic landscape is calculated. The three-dimensional color quantity is used to optimize the three-dimensional image of the urban landscape. The detailed steps are as follows: It is assumed that represents the color leafed plants. M(nk) represents the characteristics of landscape design as a whole and contour. Based on the physiological and ecological characteristics of color plants obtained above, we use the formula (7) to describe the three Informatica 43 (2019) 325–330 327 dimensional spatial structure index of urban humanistic landscape. (..) .(....)...(km) (....).... = ×....(..,..) (7) ..(....)...(....) In the formula, (kJ) represents the three dimensional Green amount, and the ..(....) represents the crown diameter. It is assumed that represents the non three-dimensional color leafed plants. ..(....) represents three-dimensional color leafed plants. The formula (8) is used to calculate the amount of color leafed plants: ..(....)*..(....) ....(..)= (..)×..(..,..)...(..,..) (8) (....).... In the formula, R(k,l) represents the color of the actual plant, v(i,l) represents the classification of color leafed plants by leaf color characteristics of the situation. It is assumed that stands for the three dimensional color sum of common leaf, bicolor and spotted leaves. m(k,l) represents the crown of color plants, and represents the three dimensional color sum of autumn leaf, new leaf color, common leaf, bicolor, and leaf color. The m(k,l) is defined as the amount of color leafed plants in spring, and the is defined as the color leafed plants in summer and winter. The formula (9) was used to calculate the three-dimensional color conversion rate of urban landscape plants. ..(..,..)·.... ....(..) ......(..,..)= ....(..)·(9) ..(..,..) ..(....) In the formula, nf(b) represents all the colorful plants of the urban humanistic landscape, and ..(....) represents the sum of the three-dimensional color of all the colorful plants of the city's humanistic landscape [5]. It is assumed that l(b,m) represents the sum of the three dimensional Green quantities of non-colored plants. ..(..,..) represents the difference in the contrast of three dimensional coloring rate in spring. The model of the three-dimensional image optimization of the urban landscape plant landscape was established by using the type (10). ..(..,..) ..(....)= ×......(..,..)..* (10) ..(..,..) In the formula, M* represents the number of different phase accumulation value of colorful plants. Simulation results and analysis In order to prove the validity of the proposed 3D image optimization design model of urban humanities landscape, an experiment is needed. In the environment of Matlab, a simulation platform for three-dimensional image optimization of urban humanities landscape is built. The data of humanistic landscape resources in urban residential areas from May 2016 to October 2017 are the experimental data [6]. 4.1 Different models of landscape design color and three-dimensional color comparison This paper uses the model proposed in this paper and the literature model to carry out the experiment of the landscape design of urban landscape plants. The three-dimensional color and three-dimensional color rate of plant color landscape design were compared between the 2 different models. The comparison results are shown in figures 2 and 3. From the analysis of figures 3 and 4, it can be concluded that the color and rate of color rendering of urban humanistic landscape plants is better than that of the literature model. This is mainly because that when we use this model to design the 3D image of plant landscape in urban landscape, we first get the basic conditions of rational collocation of plants in landscape based on the principle of highlighting urban characteristics and ecology. According to the needs of the landscape theme, the colorful landscape elements of the landscape plants are obtained, and the seasonal variation of color plants is obtained. The physiological and ecological characteristics of the color plants are given, which ensures the color rate and three-dimensional color of the three-dimensional image design of the urban landscape plant landscape [7]. 4.2 Comparison of the effectiveness of different models for landscape design The model in this paper and literature model were used to design the experiment respectively. The stability(%) and efficiency(%) of the 2 different models for the rational allocation of urban cultural landscape were compared [8]. The comparison results were used to measure the overall effectiveness of the 2 different models of urban humanistic landscape design. The results were shown as figure 5 and figure 6. From the analysis of Figure 5 and figure 6, it can be concluded that the overall superiority of the urban humanistic landscape design using this model is better than the model in the literature. In this paper, the three-dimensional color quantity concept of the urban humanistic landscape is designed, and the three-dimensional color quantity of the urban humanistic landscape is calculated. The three-dimensional image of urban humanistic landscape is optimized by three-dimensional color, which ensures the overall superiority of the landscape design. W. Wu et al. The trend of humanistic landscape design in urban residential areas With the accelerated process of urbanization in China, the design of humanistic landscape in urban residential areas is also developing. City residential area cultural landscape is more geared towards life, which meets the needs of the residents in the direction of development. It tries to improve the quality of the entire urban residential area with the most shared, cultural and artistic humanistic landscape. As shown in Figure 7. (1) The development of the shared cultural landscape When designing the residential landscape, we should take into account the needs of all users, and design a shared cultural landscape, so that every resident can enjoy and enjoy these cultural landscapes together. (2) Cultural development of the humanistic landscape In the design of the humanistic landscape, it is necessary to integrate more history and culture on the basis of the natural environment, so as to make the cultural landscape richer. Natural landscapes are naturally formed, while humanistic landscapes are designed by human beings, and human history and cultural development are the source and foundation of design. Therefore, humanistic landscape is actually a human cultural landscape. Because of the unique cultural background and historical background, the residents can feel the edification of art and historical culture when they appreciate the cultural landscape. Therefore, in the process of development, the pursuit of culture is an important direction in the development of humanistic landscape. (3) The artistic development of the humanistic landscape (4) The good development of the humanities landscape In order to attract household occupancy to a greater extent, many residential areas introduce the water system to the construction of humanistic landscape. It can realize the scientific and rational arrangement of each layout, which can realize the harmonious development of human, environment and society. Informatica 43 (2019) 325–330 329 With the development of social economy, people’s pursuit of art is higher. Therefore, in the process of design, humanistic landscape needs to satisfy people's artistic pursuit and aesthetic need. The humanistic landscape not only maintains the natural ecology, but also has the artistic beauty [9]. 6 Conclusion With the continuous improvement of people's living standards, the demand for the living environment is becoming higher and higher. Therefore, the design of the green landscape in the residential area can not only meet the needs of the residents for the green space, but also improve the beauty and landscape culture of the residential area. More importantly, it can play a very important role in the greening of the city and the improvement of the ecological environment. We should pay attention to the ancient and emerging disciplines of the cultural landscape. This design accords with the modern and practical function of the poetic urban garden new space, and creates the beautiful landscape human settlement ecological environment. This should be the goal of all planners. 7 References [1] Jaeger J A G, Bertiller R, Schwick C, et al. Urban permeation of landscapes and sprawl per capita: New measures of urban sprawl[J]. Ecological Indicators, 2010, 10(2):427-441. https://doi.org/10.1016/j.ecolind.2009.07.010 [2] Smallbone L T, Luck G W, Wassens S. Anuran species in urban landscapes: Relationships with biophysical, built environment and socio-economic factors[J]. Landscape & Urban Planning, 2011, 101(1):43-51. https://doi.org/10.1016/j.landurbplan.2011.01.002 [3] Mckinney R A, Raposa K B, Cournoyer R M. Wetlands as habitat in urbanizing landscapes: Patterns of bird abundance and occupancy[J]. Landscape & Urban Planning, 2011, 100(1–2):144­152. https://doi.org/10.1016/j.landurbplan.2010.11.015 [4] Mcdonald R I, Forman R T T, Kareiva P, et al. Urban effects, distance, and protected areas in an urbanizing world[J]. Landscape & Urban Planning, 2009, 93(1):63-75. [5] https://doi.org/10.1016/j.landurbplan.2009.06.002 [6] Peterson M N, Thurmond B, Mchale M, et al. Predicting native plant landscaping preferences in urban areas[J]. Sustainable Cities & Society, 2012, 5(1):70-76. https://doi.org/10.1016/j.scs.2012.05.007 [7] Devitt D, Morris R, Gianquinto G P, et al. Sustainable water use in urban landscapes in the 21st century: a Las Vegas perspective.[J]. ActaHorticulturae, 2010, 881(881):483-486. https://doi.org/10.17660/ActaHortic.2010.881.77 [8] TarekRashed, JohnWeeks. Assessing vulnerability to earthquake hazards through spatial multicriteria analysis of urban areas[J]. International Journal of Geographical Information Systems, 2003, 17(6):547-576. https://doi.org/10.1080/1365881031000114071 [9] Kilpatrick H J. Effects of archery hunting on movement and activity of female white-tailed deer in an urban landscape. WildlSocBull[J]. Wildlife Society Bulletin, 1999, 27(2):433-440. [10] Endterwada J, Kurtzman J, Keenan S P, et al. Situational Waste in Landscape Watering: Residential and Business Water Use in an Urban Utah Community[M]. American Water Resources Association, 2008. [11] https://doi.org/10.1111/j.1752-1688.2008.00190.x Research on Dance Teaching Mode Based on Flipped Classroom in the Internet +Age Fubo Ma Art Academy, Northeast Agricultural University,China E-mail:fuboma@126.com, http://www.neau.edu.cn Chang Guo College of Engineering, Northeast Agricultural University,China E-mail:ceinuo@126.com, http://www.neau.edu.cn Keywords: Internet, flipped classroom, dance teaching mode Received: July 15, 2019 The development of Internet technology has injected new impetus into the reform of education, and also posed new challenges to the traditional teaching mode. In recent years, as an extension of educational informationization, the flipped classroom has developed rapidly, and academic attention has increased year by year. At present, the flipped classroom in Colleges and universities in China has penetrated into many disciplines. Dance teaching should conform to the development of the times, optimize the teaching mode of dance education, open up innovative consciousness, and form a multi-channel teaching pattern. On the basis of discussing the theoretical and practical research on the mode of flipping classroom teaching in the dance teaching, this paper puts forward and designs the structure map of the pattern of the teaching mode of the flipped class in the ordinary university, and carries out a 4 month teaching experiment for the dance students. Through the analysis and verification of the data from the investigation and feedback, the conclusion is that the teaching mode of the dance flipping classroom has significant effect on improving the students’ basic dance quality, dance performance and autonomous learning ability. Povzetek: V prispevku je opisana uporaba umetne inteligence za izboljšanje učenja plesa. Introduction In twenty-first century, the rapid development of the information technology revolution and the continuous upgrading ofthe technology are slowlychangingpeople’s life, work and communication ways. It also brings opportunities for a series of changes in education and provides technical feasibility for the wide spread of classroom teaching. The current digital campus and the Internet of things have laid the material foundation for the flipping classroom. Mobile terminals provide technical tools for flipped classroom learning. At present, the advanced educational resources, such as micro video, Khan College and electronic desk, have provided convenient access to the implementation of the overturned classroom. With the continuous upgrade of technology, more and more information tools and methods will emerge to make learning easier and faster. To sum up, the development of modern information technology is the necessary condition for the smooth expansion of the flipped class, and the implementation of the information technology course cannot be separated from digital equipment. The development of technology has promoted the development of information technology courses and flip-flop classes, and the research on the teaching mode of junior dance based on flip-flop classes is essential. In the application of the overturned classroom teaching mode in the dance class in college sports, Zhang Qin (2016) concludes that the flipped classroom teaching mode is applied to the teaching of dance and dance courses in colleges and universities. It needs to be carried out in four aspects: the review of the content of video teaching before class + basic dance skills + effective autonomy and cooperation Learning + effective evaluation and careful ending design [1]. 2 The construction of dance teaching mode based on flipped classroom in the Internet age 2.1 The application of the flipped class in the Internet age From the implementation to the operation, the overturned class can be divided into four parts. The first is the preparatory stage of teaching, the second is the teaching memory and the understanding, the third is the implementation and the analysis stage, the last one is the comprehensive evaluation stage. The most important feature of this model is the reversal of roles. At the preparatory stage, the teacher is the protagonist, and in the understanding stage, the student becomes the protagonist in the classroom. At the stage of application analysis, students and teachers are both the protagonists. The last stage is mainly teachers scoring, which can be done by group [2]. 2.2 Design of dance teaching mode based on flipped classroom Because of the specialty of dance teaching, body movement is the expression form of the practitioner [3]. The practitioner needs to master the technical movements and practice a lot under the guidance of the teacher to form the dance skills. In the construction of the reversed classroom teaching mode in dance education, it is necessary to focus on the needs of students, so as to construct a more suitable dance education teaching mode in Colleges and universities. The structure diagram consists of three stages (Figure 1). They are pre-class preparation of classroom resources, in-class knowledge internalization and content upgrade, and after-class consolidation and improvement. According to the curriculum requirements of dance discipline, theoretical knowledge is understood and mastered. The process of the design of the structure drawing emphasizes the importance of communication. The process of communication and feedback is highlighted at the beginning of class. It breaks through innovation in the way of evaluation, and makes full use of teaching resources and communication platform. It highlights the advantages of the flipped classroom and combines the characteristics of dance education to provide reference for the reform of dance education [4]. Figure 1: Structure of college dance teaching mode based on flipped classroom. In the implementation steps of dance flipping classroom, it is designed in three stages (Figure 2).In the pre class stage, the teaching content should be cut in modules to set up the resources in various forms. It should be intuitionistic and effective. Students should conduct self-identification in the form of online testing or feedback. In this structure, the importance of feedback communication is emphasized, and new ideas are provided in the production of teaching content. The division of labor between teachers and students is clear and effective, and the design of teaching platform is more humanized. The way of online evaluation helps teachers and students to summarize and reflect, to further optimize the teachingplan and tosupport the students’ learning and practice [5]. F. Maet al. Figure 2: The steps of the dance flipping class. The role of the teaching mode is mainly embodied in the teaching effect, which also reflects the current situation of teaching in China. The biggest feature of the flipping class is that it can improve the students’ interest in learning and attract the attention of the students in class so as to improve their achievements and improve the teaching efficiency [6]. 3 Practice research of dance teaching based on flipped classroom 3.1 Research objects and methods In this paper, the application effect of flipped class in the teaching of dance in Colleges and universities is studied. A total of 120 students in the last semester of 2017-2018 year’s dance courses were tested for a total of nine weeks of teaching experiments. The experimental class and the control class had a weekly class of 90 minutes each. The two classes consist mainly of two parts. One is the basic pace of dance, and the other is a small group of teachers. According to different teaching methods and reasonable arrangement order, the key points and difficulties are noted [7]. Evaluation method Account for the proportion of achievement s Evaluation subject Attendance and feedback video frequency 30% Teacher evaluation Classroom participation performance 20% Teacher evaluation, group evaluation, self-evaluation Group presentation 20% Teacher evaluation, group evaluation, self-evaluation Table 1: Flipped classroom process evaluation. As can be seen from the evaluation table, the main evaluation methods are the number of students in class and Research on Dance Teaching Mode... the number of feedback videos, classroom performance and group presentation. The number of classes and the frequency of feedback are evaluated by teachers. Class participation is mainly evaluated by students, groups and teachers. Grading project Scoring index Scoring body Stance (5 points) Basic posture grip (5 points) (20 points) back (5 points) head (5 points) Integrity (10 points) Combined action fluency (10 points) (40 points) direction (10 points) Teacher trajectory (10 points) Music performance Music rhythm (30 points) (30 points) Facial expression Facial expression and clothing (5 points) (10 points) Clothing (5 points) Table 2: Flipping class final skill test score. The result of the evaluation is mainly the skill test at the end of the term, and the content of the examination is also the proportion of the score. It is mainly obtained by experts through questionnaires and data analysis. The main contents of the grading are basic posture, combined movements, music rhythm, facial expressions and clothing. The full score is one hundred points. The basic position is twenty points, while standing and holding posture and back and head are five points separately. The composition movement was forty points, the movement integrity, fluency, direction, track accounted for the ten points separately [8]. Musical rhythm has thirty points, facial expressions and clothing account for ten points. At the end of the final assessment, we invited three dance teachers from our school to carry out the score, which was carried out with a fair and strict attitude. The average score of three teachers are taken as the final result. 3.2 Pre-test and post test data analysis 3.2.1 Pre-test results and analysis Before the experiment, the level of dance technology is measured. A basic step in dance is the main content of the test. At the same time, the basic situation of dance has also been investigated through questionnaires. In this way, we can get the cognition of the control class and experimental class to the dance. According to table 3, the students in the two classes have not been trained professionally. Only 3.3% of the students have been in contact with the dance before, and the other students have not been in contact with them. Therefore, the cognition of the two classes is zero. So the two classes have professional knowledge and technical mastery is relatively low, both classes are on the same starting line [9]. Informatica 43 (2019) 331–336 333 Is there any professional training before Have you ever been in contact with a dance before Yes No Yes No Experi mental class Number of people 0 60 2 58 Percentage 0 100% 3.3% 96.7% Control class Number of people 0 60 2 58 Percentage 0 100% 3.3% 96.7% Table 3: Investigation and statistics on basic situation of dance in experimental class and control class. This article mainly takes the exercise attitude scale as the main weighing work. This table is mainly for young students, and there are eight main indicators. They are behavior attitude, behavior cognition, behavior habit, behavior intention, objective attitude, emotional experience, behavior control and subjective standard. Each index is formed by a different number of items. There are seventy entries, and the higher the score is, the better it is. In order to get a better understanding of the attitude of the two classes, the two classes were measured before the experiment, and the results of the measurements were found through table 4. Through table 4, we can learn that the two classes have more than 0.05 P values in behavior attitude, target attitude, behavior cognition, habit, intention, emotional experience, behavior control and subjective standards. This shows that the two classes are not very different in these eight aspects. The exercises of the two classes are kept at the same level so as to meet the requirements of the experiment. In order to get a detailed understanding of students’ interest in dance lessons, a questionnaire survey was conducted before the experiment. According to the statistics, only 5% and 7% of the students in the two classes are very interested in learning dance, only 15% and 20% are interested. There are only 37% and 30% of the students with general interest. Those who were not interested accounted for 40% and 37%, while those who did not like it accounted for 3% and 5% respectively. This also means that students are not interested in learning dance [10]. Experimen tal class (n=60) Control class (n=60) X±S X±S t value P value Behavior attitude 28.36±6.59 27.19±4.20 0.914 0.361 Objective attitude 40.01±7.36 41.02±5.38 0.775 0.436 Behavior cognition 24.49±4.76 25.48±5.08 1.084 0.282 Behavior habits 35.54±8.73 33.59±5.91 1.152 0.253 Behaviora l intention 24.34±5.43 25.51±5.34 1.106 0.324 Emotional experienc e 33.35±6.32 34.88±6.27 1.102 0.264 Behaviora l control 23.03±5.46 21.47±4.52 1.421 0.164 Subjective criteria 20.45±5.38 22.01±6.32 1.153 0.253 Table 4: Independent sample T test results of exercise attitude before and after test class. In order to investigate the abilities of the two classes, a questionnaire survey was conducted before the experiment. According to the survey, it is found that the two classes have very little study before class, and most of them do not have to consolidate their study after class. Only a few students choose to listen to the teacher when they encounter something they don’t understand in class. This also indicates that the students’ learning situation is not very good. 3.2.2 Post test results and analysis The degree of acceptance is mainly observed from the object being carried out, and whether the students can accept the dance class in the flipped class, mainly based on the degree ofthe students’ preference. In order to get a clear understanding of students’ acceptance, especially after the end of the experiment, a questionnaire survey was conducted among the students in the experimental class [11]. Figure 3: An analysis of the acceptance of the flipped classroom teaching. According to figure 3, it can be found that the person who likes it occupies 57%, more than half of the people like it. People who like it occupy 20%, and those who have the general interests have 13%.There are 10% people who don’t like it. The people who like todance stilloccupythe vast majority. For the form of flip classroom teaching, students are still willing to accept it, especially the dance teaching under the flipped classroom [12]. First, a simple analysis of the results of the two classes is made from the basic positions, the movements of the combination, the rhythm of the music and the facial expression. According to table 5, we can learn that a=0.05 is a basic test standard, while the students in the two classes have certain differences in the basic posture, the combination of the movements and the rhythm of the music, which also represents a certain effect in the flipped class at this time. The students in the two classes did not change in facial expression and clothing. According to table 6, it can be found that the average values of the students in the two classes are quite different, and the P value is less than 0.05.The result of the comparison between the two classes shows that the students in the experimental class are better than those in the control class. During the period after the end of the teaching experiment, we conducted an independent sample T test on two classes (Table 7). According to table 7, we can well observe the behavior attitude of the students in the experimental class and the control class, as well as the target attitude and the behavior habits. The six aspects of the P value are less than 0.05, and this also represents the gap between the students of the two classes in these six aspects. The P value was higher than 0.05 in cognition and behavior control, which also showed that there was no significant difference between the two classes. Experimental class (score) Control class (score) P value X±S X±S Basic attitude achievement 17.6±1.605 15.01±2.492 0.005 Combined action 36.05±2.182 32.98±3.267 0003 rhythm of music 30.02±1.237 26.47±2.684 0.004 Facial expression and clothing 9.41±0.482 9.16±0.369 0.250 Table 5: Analysis of four indicators of technical performance in the experimental class and the control class Experimental class (score) Control class (score) P value X±S X±S Technical test results 93±4.602 86.42±6.387 0.008 Table 6: Total score analysis of final examination in experimental class and control class After a period of learning, the dance learning in flipped classroom has achieved certain results. At the same time, there were questionnaires for students in the experimental class as well as in the control class. In the experimental class, about 60% of the students were interested in dance, and 19% of them had increased interest. Only 10% of them were only generally improved, and 5% still felt no interest. While 5% of the control class showed a significant improvement in their interest in dancing, 10% showed a general improvement, and 79% showed no change. Through these data, we can also know that flipped classroom has played a very positive role in enhancing students’ learning [13]. 3.3 Implementation effect analysis Through the experimental data, we can see that the dance skills of the experimental class are higher than those of the control class. And the exercise attitude of the experimental class is better than the control class, and this also shows that the effect of the teaching in the flipped class has already achieved certain effect, so it is still very applicable in the University. Experiment al class (n=60) Control class (n=60) X±S X±S t value P value Behavior attitude 26.01±5.08 25.53±5.29 1.326 0.043 Objectiv e attitude 44.89±5.98 41.69±5.79 2.302 0.026 Behavior cognition 24.72±4.35 26.01±5.31 3.379 0.538 Behavior habits 37.01±5.68 33.29±5.36 2.702 0.009 Behavior al intention 37.06±6.06 33.06±8.32 2.245 0.017 Emotion al experien ce 25.58±6.70 28.87±6.72 -2.226 0.011 Behavior al control 22.98±5.51 21.43±4.64 1.412 0.164 Subjectiv e criteria 25.25±5.23 22.72±5.31 2.134 0.034 Table 7: Independent sample T test results of exercise attitude of experimental class and control class Before class, teachers can make some micro-videos for teaching, which helps students consolidate their knowledge points after class. And students can do exercises repeatedly after class. In addition, students can learn in groups in the classroom, which is also conducive to improving students’ learning ability [14]. The arrival of flipped classes makes students feel very fresh, which to a certain extent also increases students’ interest inlearning. The students’ self-learning ability and cooperative learning ability are all exercised in all aspects. Their subjectivity has been brought into play, and the efficiency and quality of the teaching have been improved. 4 Conclusion Based on the exploration and research of the new education model, the author recognizes that the new education mode in the overturning classroom is the trend of future development. The mode of flipping class extends the students’learningprocesstothe class.The studentsget the information they need quickly through the network means. Compared with the traditional class, the flipping classroom teaching mode is more helpful to stimulate the students’ interest inlearning and mobilize the enthusiasm, initiative and creativity of the students’ study. But this is not the only teaching mode. Because the educational situation in some places may not be appropriate, it cannot be implemented in a comprehensive way. Some developed areas can be decided according to the actual situation. For some areas with relatively backward economic conditions, careful consideration should be given. 5 References [1] Smith S, Brown D, Purnell E, et al. ‘Flipping’ the Postgraduate Classroom: Supporting the Student Experience[M]// Global Innovation of Teaching and Learning in Higher Education. Springer International Publishing, 2015:295-315. https://doi.org/10.1007/978-3-319-10482-9_18 [2] Lin Y N, Hsia L H, Sung M Y, et al. Effects of integrating mobile technology-assisted peer assessment into flipped learning on students’ dance skills and self-efficacy[J]. Interactive Learning Environments, (1), pp.1-16, 2018. [3] Ting H H, Brito J P, Montori V M. Shared decision making: science and action[J]. Circulation Cardiovascular Quality & Outcomes, 7(2), pp.323, 2014. https://doi.org/10.1161/CIRCOUTCOMES.113.000 288 [4] Fevolden A M, Tmte C E. How Information and Communication Technology Is Shaping Higher Education[M]// The Palgrave International Handbook of Higher Education Policy and Governance. Palgrave Macmillan UK, 2015. https://doi.org/10.1007/978-1-137-45617-5_19 [5] Feng J Y, Li W Q. Practice and Research on Flipped Classroom Teaching based on MobileInternet[J]. Building Technology Development, 2017. [6] Wen-Ying L U, Xian Peihua University. Research on Teacher-Student Interaction in Flipped Classroom Teaching Mode Based on MOOCs[J]. Journal of Tianjin Sino-German Vocational Technical College, 2016. [7] Liu X. Research on "Flipped Classroom" Teaching Mode Based on Microlecture[J]. Shipbuilding Vocational Education, 2017. [8] Fang F. Research on the Teaching Mode of Specialized Skills Courses Based on the Idea of Flipped Classroom:A Case Study of "Determination of Total Hardness of Water" in the School-based Textbook “Routine Detection”[J]. Science Education Article Collects, 2017. [9] Zhang X, Computer S O. The teaching research of the dynamic website construction course based on the teaching mode of MOOC + flipped classroom teaching mode[J]. Journal of Science of Teachers College & University, 2017. [10] Luo P, Xia-Fu L V, Min L I. The Research on Flipped Classroom Teaching Mode[J]. Education Teaching Forum, 2017. [11] Liu J, Wang H. Research and Practice of the Teaching of Body Shaping Course Based on the Concept of Flipped Classroom[J]. Shanxi Science & Technology, 2016. [12] Zhao J, Guan M, Jiang L H. Research on Operations Teaching Mode Reform under Modular Teaching— —Based on Flipped Classroom[J]Journal of Anhui Science & Technology University, 2016. F. Maet al. [13] Siobhan B. Mitchell,Anne M. Haase,Sean P. Cumming,Robert M. Malina. Understanding growth and maturation in the context of ballet: a biocultural approach[J]. Research in Dance Education. 2017(3) https://doi.org/10.1080/14647893.2017.1387525 [14] Janet Withall,Anne M. Haase,Nicola E. Walsh,Anita Young,Fiona Cramp. Physical activity engagement in early rheumatoid arthritis: a qualitative study to inform intervention development[J]. Physiotherapy. 2016(3) [15] https://doi.org/10.1016/j.physio.2015.07.002 Research on Development Mode of Intelligent Rural Tourism under Digital Background Chunlin Zhou School of Tourism and Event, Henan University of Economics and Law, China E-mail: zhouchunlin@huel.edu.cn http://www.huel.edu.cn Keywords: digital background; smart village; tourism development Received: July 15, 2019 Intelligent rural tourism rapidly emerged under the digital background in China after the reform and development. Driven by the theme of China tourism administration's smart rural tourism year under digital background, the digital background smart rural tourism market has entered a period of vigorous development, accelerating the integration of tourism and agriculture, with forestry and other related industries. Since the late 1990s, China's rural areas have experienced four stages of spontaneous development, quantity expansion, standardized development and quality improvement. After more than 20 years' active exploration, China has established a development path with Chinese characteristics and a smart rural tourism model under digital background. Povzetek: Opisan je razvoj inteligentnega kmečkega turizma na Kitajskem z metodami UI. 1 Introduction Intelligent rural tourism rapidly emerged under the digital tourism reception capacity and income in China as shown background in China after the reform and development. in Table 1. Driven by the theme of China tourism administration's smart rural tourism year under digital background, the digital background smart rural tourism market has entered a period of vigorous development, accelerating the integration of tourism and agriculture, with forestry and other related industries and industries. Since the late 1990s, China's rural areas have experienced four stages of spontaneous development, quantity expansion, standardized development and quality improvement [1] [2]. After more than 20 years' active exploration, China has established a development path with Chinese Table 1: Intelligent rural tourism recipients, employees characteristics and a smart rural tourism model under and income in China. digital background. Time Smart rural tourism reception (ten Smart rural tourism employees (ten Smart rural tourism income 2010 1404 62 251 2011 157 64 304 2012 162 68 359 2013 168 78 376 An investigation into the smart rural tourism in China found that the digital background smart rural tourism 2 Research on the development products presented traditional singleness and lack of diversity. The overall level stayed at a low level of mode of intelligent rural tourism offering accommodation, agricultural products, parties under digital background and souvenirs [4]. The number of smart rural tourism and 2.1 Analysis of the development status of digital rural intelligent tourism in China In 2009, in order to upgrade China's smart rural tourism under digital background, the National Tourism Administration introduced a variety of new digital background smart rural tourism on the basis of provincial smart rural tourism with the digital background. The new concept of smart rural tourism is of great significance for enhancing the scale and quality of digital rural tourism [3]. In recent years, the digitized background of smart rural the number of receptionists in China in recent years is shown in Figure 1. 2.2 Digital background definition of intelligent rural tourism development model This paper defines the digital background smart rural tourism development model as the chain-type relationship formed by certain departments supporting digital rural intelligence tourism activities through certain economic relationships [5]. The smart rural tourism development model comprises of a horizontal extension, vertical associated function chain, product chain and cultural value chain. It appears as a network structure consisting of horizontal cooperation and vertical supply and demand. Its expression is as follows: ..=..(..,..,..,..) (1) In this concept, there are the two points to be explained. For one thing, some departments on the digital background smart rural tourism development model are horizontally cooperative. Each industry in the smart rural tourism development model can directly provide tourism products for tourists, but they need to work together to provide tourists with complete tourism product. For another, if the collaboration is not effective, the poor performance of one of the industries will weaken the development pattern of the entire digital background smart rural tourism [6]. The change relationship is as shown in formula (2) as follows: (2) 2.3 Digital background characteristics of intelligent rural tourism development model (1) Reticulation structure The traditional mode of manufacturing rural tourism development is based on the vertical industrial linkage of the product process division. Products are processed through upstream, middle, and downstream enterprises and finally presented to consumers [7]. The background of the digital background smart rural tourism development model includes not only the vertical supply and demand relationship, but also the horizontal collaboration relationship. Its mathematical expression is as follows: ..=(..,..1,..2,....) (3) Digital background intelligent rural tourism products have comprehensive features [8]. Tourists enjoy a range of tourism services from the time they leave their place of residence for the digital background smart rural tourism destinations. In other words, different from the final products of the rural tourism development model for the consumer, the various departments on the digital background smart rural tourism development model can directly provide tourism consumers with tourism products. C. Zhou However, each department can only provide part of the product [9]. If tourists are to be able to experience the full digital background smart rural tourism products, they need effective collaboration and cooperation among all sectors of the rural tourism development model. The main coordination method is shown in formula (4) as follows: .. ....+.. =. () (4) ..(....+..|....)..=1 ....+.. (2) The characteristics of digital background intelligent rural tourism The rural nature is the essential attribute of the digital background intelligent rural tourism. The reality includes the rural nature of tourism resources, tourism products, tourism market and tourism benefits; the rural nature can be divided into rural culture and rural landscape [10]. The digital background essence of the rural tourism is the rural culture. The rural culture with nationality, history and region is the essential attribute of the digital background intelligent rural tourism. Rural culture as a concrete description is the property of the digital background intelligent rural tourism. Its mathematical expression is as follows: ..=...,................ (5) The determination of core value of the digital background smart rural tourism development model is conducive to a clear direction for development. As to the core of the general tourism rural tourism development model, the traditional view is that travel agencies are the hubs linking the six major elements of tourism, while under the digital background, the reconstruction of the value chain of the tourism industry aims at reconfirming the core node of the chain, and that is, establishing a tourism industry value chain model with tourism sites and tourist attractions as the core [11]. The core of tourism and rural tourism development model is tourism and tourist attractions, and the tour experience of tourism sites and tourist attractions. The core value is tourism and rural tourism development model. However, the digital background smart rural tourism is a more special form of tourism. the mathematical expressions of the four traditional forms of digital rural tourism in China and eight new forms of business are as follows: ...... =(..,..) (6) People choose smart rural tourism for various reasons, some to taste rural food, some to experience agricultural activities, and some to enjoy the rural scenery. That is to say, people choose smart rural tourism for the specific rural cultural atmosphere. It is a kind of cultural yearning for rural food culture, agricultural culture, and rural landscape culture. Therefore, this paper holds that the core value of the digitalized background intelligent rural tourism development model should be rural cultural experience. The value system centering on the intelligent rural tourism cultural experience is the basis for realizing the value-added rural tourism development model [12]. Culture is a description of a special way of life, and then the digital background smart rural tourism is a description of the special rural lifestyle. The rural culture includes rural food culture, rural living culture, rural landscape culture, agricultural culture, rural product culture, rural entertainment in terms of culture and so on, and these aspects are all examples of rural special lifestyles as shown in Table 2: Rural cultural experience manifestation Performance description Rural food culture experience Including rural eating habits, methods, and allegorical relations with literature and art Rural characteristic folk customs experience Including the selection of rural residential sites, construction techniques, architectural structures, spatial layout, and aesthetic ideas, religious concepts, etc. Rural landscape / pastoral scenery experience Different from the comprehensive performance of various phenomena such as the humanities, society, and nature of Rural entertainment culture / farming activities experience Including picking, fishing and other farming or recreational activities. Table 2: Multi-dimensional performance of core values of smart rural tourism development model with digital background. (3) The characteristics of China's intelligent rural tourism evaluation There are many perspectives on intelligent rural tourism evaluation. For example, the stability of rural tourism development model can be evaluated from the government's point of view. The cooperation of rural tourism development model can be evaluated from the perspective of the enterprise [13]. With tourism development model for performance evaluation, this paper chooses to evaluate the rural tourism development model of China's digital background smart rural tourism from the perspective of tourists. There are two main advantages in selecting tourist perspectives: firstly, assessing the advantages and disadvantages of the rural tourism development model from the tourists' perspective is closer to the target market; secondly, the use of questionnaires on the tourists is for the measurement of rural tourism development model, and the result is more objective and easier to measure. This paper selected digital background smart rural tourism catering, digital background smart rural tourism accommodation, digital background smart rural tourism products, digital background smart rural tourism scenic spots, digital background smart rural tourism traffic, digital background smart rural tourism entertainment activities for the project indicator layer, and there are several factor layers under the indicator layer, as shown in formula (7) as follows: ..(....)=..(....|......(....)) (7) 2.4 Data analysis of smart rural tourism development model under digital background (1) Questionnaire design The questionnaire in this paper is divided into two parts: firstly, the demographics and travel characteristics of tourists, including the demographic characteristics of tourists and tourists' travel characteristics. Secondly, the development model and importance. Development model survey factors use the five-point scale as the evaluation criteria for index evaluation. The survey objectives are: development model and importance. This paper focuses on field surveys of smart rural tourism sites with digital backgrounds in China. 288 questionnaires were distributed and 268 valid questionnaires were collected. (2) Reliability analysis A questionnaire with good reliability shows good stability and consistency. Reliability analysis is the analysis method that obtains the reliability of the questionnaire through the evaluation method, and uses the reliability to evaluate whether the questionnaire has good reliability and stability [14]. In general, the same question in the questionnaires of different subjects tends to be consistent, indicating that there is no large sample-to-sample error in the questionnaire survey and the reliability of the questionnaire is higher, and the reliability is lower. Reliability generally uses . coefficient as its evaluation index. The coefficient is generally between 0 and 1 [15]. The higher the reliability, the closer of . coefficient is to 1, which means that the more stable the questionnaire, and the mathematical expression is as follows: ....) ..=(..,..,h,h..,...,h(8) Under normal circumstances, . coefficient is above 0.70, which indicates that the questionnaire is of good reliability and can be used for the next analysis. If . coefficient is lower than 0.35, and the reliability is low, the stability of the questionnaire is insufficient, and it cannot be used. Among them, the reliability of the questionnaire is acceptable [16]. In this study, SPSS tool was used to analyze the reliability of the questionnaires used, and their consistency was checked to ensure the quality and credibility of the questionnaires. After inspection, the overall reliability of the questionnaire used in this survey was .=0.956, and the reliability of development model .=0.954, and the importance of reliability .=0.961, all in line with Cronbach's standards, showing good reliability of the questionnaire. (3) Development model result analysis In this survey, men accounted for 45% of the total number of samples, and women accounted for 55%, and the gender ratio is more balanced. In terms of age structure, the proportion of tourists aged 22-38 is the highest. From the point of education, undergraduates and college graduates accounted for the most. In view of residence, the majority of the tourists are Chinese. We can calculate the mean and standard deviation of the development model of 20 evaluation indicators, and sort them according to the average size. Through observation, Indicator layer Factor layer Very dissatisfied Dissatisfied General Satisfaction Very satisfied Average Rural Tourism Catering Rural features 6 9 123 95 16 3.7 Environmental hygiene 7 23 154 53 5 3.5 Price 15 45 93 72 18 3.8 Rural tourism accommodation Rural features 2 14 122 85 13 3.3 Environmental hygiene 8 48 93 72 12 3.6 Price 5 33 107 98 4 3.7 Rural tourism products Indicator layer Rural Tourism Catering Rural features 5 27 122 74 16 3.5 Price 8 64 103 53 3 2.2 Quality 7 58 124 52 2 3.4 Table 3: Ranking of development patterns of evaluation factors for intelligent rural tourism in China. it can be seen that among the three-level indicators, tourists are more satisfied with the rural characteristics of tourism accommodation, rural characteristics and tastes of tourism and catering. The secondary models of development are the quality, price, and prices of tourism and catering. Paired sample T-test is a statistical method to test whether there is a significant difference in the overall mean of two paired samples. Paired samples can be two sets of sampled data for the same variable, and can also be considered as two different aspects of a problem [17]. To conduct paired sample T test, firstly we need to find the difference between each pair of samples, and then compare the mean and the average value of the 0 relationship, with weak sample no difference and the mean value should be near 0, otherwise the sample is different [18]. In this study, the paired sample T test was used to analyze the difference between the importance of the questionnaire and the development model (the mutual comparison under the same indicator). If the difference was not significant, the importance of the indicator and the development model were examined. The performance of the entry is different, but conversely indicates significant, that is, and there is a certain distinction between the evaluation of the importance and the development model [19]. Conclusion The rural tourism industry chain is different from the general manufacturing industry chain. This paper defines it as a chain-type relationship that supports various sectors of rural tourism activities through certain economic relationships. [20] In the rural tourism industry chain, there are horizontal links between industries, and various industries face consumers at the same time; any individual link in the rural tourism industry chain can directly provide tourism products for rural tourists, but they cannot provide complete tourism products. The rural tourism industry chain is different from the general industrial chain, and its chain core is diversified, and the core value of the rural tourism industry chain is the experience of rural culture. The digital background intelligent rural tourism development model core value should be rural cultural experience. [21]The value system of the rural tourism development model centering on the digital background intelligent rural tourism cultural experience is the basis for realizing the value-added of the rural tourism development model.. 4 Acknowledgement The research in this paper was supported by Henan Science and Technology Research Program (International Science and Technology Cooperation Area) Project in 2015: Research on Safety Management and Technology for Large-scale Festival Events (NO. 152102410043), and Henan Science and Technology Department Project in 2016: Innovative Research on Safety Management of Large-scale Festival Events (NO. 162400410023).. 5 References [1] Zheng Y, Li Q L. Research on innovation of rural tourism development mode under new situation.[J]. Journal of Anhui Agricultural Sciences, 2011. [2] Wei H U. Research on Development Mode of Rural Tourism Based on Experience[J]. Journal of Anhui Agricultural Sciences, 2010. [3] Huang Z, Lin L U, Qin S U, et al. Research and development of rural tourism under the background of new urbanization: Theoretical reflection and breakthrough of predicament[J]. Geographical Research, 2015. [4] Wang H R, Han F L. On the Promotion of Rural Tourism in Heilongjiang Province Based on Intelligent Tourism Platform[J]. Journal of Qiqihar University, 2017. [5] Xi Y, Zhang Q. Research on the integration of urban and rural development in Nanjing under the background of global tourism[J]. Jiangsu Science & Technology Information, 2017. [6] Gao L A, Mei L, Liu J S, et al. Research on Rural Tourism Development under Spatial and Temporal Evolution of Tourism Flows Background——Based on Investigation Data of Tangyu Town in Lantian County[J]. Resource Development & Market, 2011. [7] Zhou G. Research on Low-carbon Rural Tourism Development in Western Minority Regions in China --from the perspective of neoinstitutional economics[J]. Interdisciplinary Journal of Contemporary Research in Business, 2013. [8] Zhang S L, Ming-Hui L I, Wang Y, et al. Research on the Countermeasures and Patterns of Rural Tourism's Development under the New Rural Construction Background[J]. Sichuan Forestry Exploration & Design, 2013. [9] Juan-Mei L I. The Research on the Vocational Education Strategies of Rural Female Labor Force under the Background of Rural Tourism Development[J]. Adult Education, 2017. [10] Sun J W. Discussion on Anhui Rural Tourism Development Model under the Background of Beautiful Countryside Construction[J]. Journal of Yangtze University, 2015. [11] Yang C, Yang C. Research on rural tourism development model under the background of precision poverty alleviation[J]. Agro Food Industry Hi Tech, 2017, 28(1):1191-1195. [12] Zhang L, Zhang M F, Tong Y. Research on the Development of Rural Tourism in Yunnan in the Background of Postmodern Tourism[J]. Value Engineering, 2014. [13] Huang S H. Research on the Interactive Development Mode of Rural Tourism and New Socialist Countryside Construction[J]. Journal of Anhui Agricultural Sciences, 2011. [14] Puhe M. Integrated Urban E-ticketing Schemes – Conflicting Objectives of Corresponding Stakeholders [J]. Transportation Research Procedia, 2014, 4:494-504. [15] Yang Y, Xia X L, Xia M Z. Design and Practice of Rural Tourism Service Platform in Jiangning District of Nanjing Based on Intelligent Software Application[J]. Journal of Nanjing Institute of Industry Technology, 2017. [16] Zhang J, Liu S S. Historical Architecture Regeneration Research Based on Rural Tourism Development[J]. Journal of Anhui Agricultural Sciences, 2015. [17] Zhang X. Research on the Interactive Mode of Agricultural and Tourism Business to Promote Rural Development Vitality:A Case Study of the Model of “Fairy Fruit Tourism in Four Seasons” in Shangyu,Zhejiang Province[J]. China Development, 2016. [18] Kaelin, Alyssa. Rural tourism development in Nepal: One village’s experience of socioeconomic Informatica 43 (2019) 337–341 341 structural transformation[J]. Social Justice Research Center Grant Awards, 2013. [19] Qiao L J, Wang J, Zhao J Y. Research on the Rural Tourism Development Mode Based on the “Tragedy of the Commons” about the Rural Tourism Resource——A Case of Hebei Province[J]. Tourism Overview, 2013. [20] Lin-Zhong S U. Research on the Farmer Work Problem and Rural Tourism Development under the Influences of Financial Crisis[J]. Journal of Anhui Agricultural Sciences, 2009. [21] Zhu G F. Research on Rural Tourism Development in Heilongjiang Province Against the Background of New Rural Construction[J]. Journal of Changchun Normal University, 2010. Study on the Multivariant Interactive Teaching Modes of College English under the Information Technology Environment Fangfang Chen School of Foreign Languages, Jinzhong University, China E-mail: fangfang_chen@163.com Keywords: Information technology environment, multivariant interactive, college english, teaching mode Received: July 15, 2019 Teaching modes can achieve twice the result with half the effort for teaching effect. The multivariant interactive teaching modes based on constructivism theory are a new kind of teaching modes supported by information technology. The multivariant interactive College English teaching modes under the information technology environment has changed the relationship between teachers and students in the classroom teaching process and the relationship between teachers and students and the teaching content. It can improve the quality of English teaching to achieve the goal of effectively improving the teaching effect of the course and improving students' ability of using English language. Based on the background of the information technology environment, this research takes the college students as the object and the multivariant interactive teaching modes as the research method and analyzes the effectiveness of the multivariant interactive teaching modes through an example. Povzetek: Opisana in testirana je multivariantna analiza učenja angleščine. Introduction The rapid development of information technology is gradually changing the traditional teaching environment and means. Foreign language educators should make full use of advanced digital technology to absorb and inherit the original teaching advantages to the maximum, making it easier for College English teaching to turn from the teacher centered to the "multivariant interactive" teaching mode between students and computers, students and students, students and teachers. Information-based "multivariant interactive" teaching is the development direction of College English teaching. Therefore, one of the main objectives of the reform of the current teaching mode is to construct a new teaching structure which can not only play the guiding role of teaching but also fully reflect the main role of the students. The multivariant interactive teaching modes under the digital technology environment have strong interactivity. It can realize interaction between students and computers, interaction between students and students, interaction between students and teachers, and interaction between students and learning content. The direction of interaction can also be one to one, one to many, or multivariant to many "multivariant interactive" [1]. 2 Related research based on the multivariant interactive teaching modes 2.1 The connotation of multivariant interactive teaching modes "Multivariate" is the meaning of multivariant elements and multivariant essential factors. "Multivariant interactive" is the process of interaction and interlinking of multivariant elements. In the course of teaching, "multivariate" refers to the teaching elements related to teaching activities, such as teachers' elements, students' elements, teaching environment elements, material conditions elements, and textbook elements and so on. Therefore, "multivariant interactive" can be defined as a process in which various elements related to teaching activities interact and influence each other. The multivariant interactive in the course of teaching is a series of teaching and learning activities, which can make use of the factors of teachers, students and teaching environment to improve the students' interest in learning actively, improve the students' learning state and effect, and achieve a high quality and efficient teaching effect [2]. 2.2 The characteristics of multivariant teaching modes (1) Integrating multivariant teaching methods. The multivariant interactive teaching modes intertwined the teaching method, teaching means, teaching content and teaching organization form into an interactive one. It changes the relatively abstract education idea into a specific operational strategy, and encourages students to feel, judge and practice and adjust their learning behaviors in an all-round way [3]. (2) An open teaching environment. The multivariant interactive teaching modes are open, and the classroom under this mode is an activity class. Before and after class, students must read a lot and grasp certain vocabularies so that they can interact effectively in class. The information technology environment also increases the interaction between students and teachers, so that students can interact with others in the virtual social situation, which helps students to play the role of emotional factors in language learning [4]. (3) An equal relationship between teachers and students. The multivariant interactive teaching mode advocates the establishment of an equal interaction between teachers and students. It respects the students' personality and experience, encourages and trains the students' spirit of independent exploration, interaction and practice and innovation, and tries to create a relaxed and harmonious interactive teaching situation. Students choose different learning methods according to their special situations and learning requirements, and actively participate in the whole interactive teaching process. (4) A variety of teaching forms. The multivariant interactive teaching modes are opposed to the stereotyped and traditional instilled teaching, which focuses on that students actively and collaboratively acquire knowledge. 3 Experimental designs of multivariant interactive teaching modes of college English under information technology environment The experimental subjects were a comprehensive university, which includes both freshmen who have just entered the university and seniors of the sophomore year. Freshmen organize classes in administrative classes, and old students independently choose courses according to the credit system management system, so students in a class may come from different professional classes. There are mainly four teachers involved in experimental teaching work, all of whom are members of the research group of this project. They have not only division of labor, but also cooperation, and are involved in listening and speaking experimental teaching [5]. This chapter intends to elaborate on the design of listening and speaking experiment teaching and the survey of teachers' teaching beliefs. (1) The research problem of the experiment 1) Whether the multivariant interactive teaching modes can play a positive role in the cultivation of students' awareness of oral communication strategic and the using frequency of communicative strategy; 2) Whether the multivariant interactive teaching modes can promote the formation of students' independent learning ability and cooperative ability; F. Chen 3) Whether the multivariant interactive teaching modes can provide students with meaningful input and output environment. (2) Experimental object The object of this study is the first-grade students of grade 16 law school in our university. They are students of two natural classes whose majors are law. Before the experiment, we calculate the average score for statistics after converting subjects learning background and English college entrance examination results into percentile, and the results are shown as shown in Table 1. The two classes who participated in the experiment used the same listening and speaking teaching materials and the independent learning network platform, that is, the "Experiencing English College English learning system", which is newly developed in the first volume of College English listening and speaking course [6]. The system is installed on the server based on campus network, and students can log in and study at any time. From table 1 and Figure 1, it can be seen that the English foundation of the subjects is basically the same. The time distribution of their extracurricular English learning every week in middle school is basically the same. The 1 class was taken as the experimental class and the 2 class was taken as the control class. Table 1: Learning background and College Entrance Examination Statistics Figure 1: The comparison situation of the subjects being studied. (3) Experimental scheme The experimental scheme is: 1) Selecting the subjects, they were divided into experimental class and control class, and their oral learning strategies, communication strategies and metacognitive strategies were investigated. 2) Assigning tasks and goals of "teachers and students, students and students, students and machine" interactive teaching for students in experimental class and developing an autonomous learning plan. 3) After the end of the experiment, the change situation of the oral learning strategies and the use of communication strategies in the experimental class and the control class were investigated through questionnaires. 4) Listening and speaking tests were conducted in both the experimental class and the control class and comparing the differences between their achievements. 5) Through the learner's autonomous learning program, the feedback table is executed, and the students' subjective evaluation of "teachers and students, students and students, students and machine" interactive teaching is understood through the students' learning diary and the students' interview. 6) The differences between autonomous learning ability and cooperative ability of the experimental class and the control class were compared. (4) Experimental tool There are three kinds of research tools used in the experiment. The first is the questionnaire, including the English learning strategy questionnaire and the cooperative feeling questionnaire. The second is audio­visual test. The third is the students' own subjective evaluation, including study diary or weekly diary and interview, and the implementation feedback form of learner's independent learning plan [7]. This research uses SPSS13.0 statistics to process data. Analysis of the Experimental Results of Multivariant Interactive Teachingin College English under Information Technology Environment 4.1 The experimental process of listening and speaking experimental teaching Listening and speaking experimental teaching started from the end of September 2016 and ended at the end of January 2017, which lasted for 12 weeks. During the first week of freshmen, the using frequency of English learning strategies, audio-visual tests and oral English learning needs analysis were conducted in the experimental class and the control class. Then the teaching plan was introduced to the experimental class, emphasizing the importance of cooperative learning and autonomous learning. In addition, an experimental class was trained in oral learning strategy, oral communication strategy and independent learning, and the content and operation of oral communication strategy and metacognitive strategy were introduced gradually. From the beginning of the second week, combined with the teaching resources (the new edition of "College English audio-visual Speaking Course 1"), the "College English learning system" has carried out the teaching practice based on autonomous learning and cooperative learning in the experimental class, which requires the students to complete each specific learning task seriously. The control class also emphasizes the significance of cooperative ability and Informatica 43 (2019) 343–347 345 autonomous learning, and encourages students to develop autonomous learning, but it does not activate oral communication strategies and strategy training. When completing each specific learning task, we advocate cooperative learning and independent learning. In January 2017, the applied situation of oral communication strategies of two classes was surveyed again. Finally, two classes took part in the audio-visual speaking ability test [8]. 4.2 Example description College English listening and speaking "teachers and students, students and students, students and machine" interaction based on autonomous learning and cooperative learning is shown in Figure 2. The solid line is strong and the dotted line is weak. Figure 2: "Teachers and students, students and students, students and machine" interaction. 4.3 Experimental results and analysis (1) Survey results of the use of oral strategy Before and after the experiment, we made statistics and analysis on the pretest and posttest of the two oral tests in the control class and the experimental class, and the survey situation of oral communication strategy using. The results are shown in Table 2 and Figure 3. Figure 3: Lateral contrast diagram. Table 3: Longitudinal statistical table of the strategy using in the control class and the experimental class. As can be seen from table 2 and figure 3, the mean values of the strategy using of the control class and the experimental class 12 weeks ago were 0.63 and 0.49 respectively, and the probability was 0.22, greater than 0.01 after the p-value test. This shows that the frequency of strategies using of the two classes is relatively low, and the use of oral strategies is almost no difference, and there is little correlation between the strategies using. After 12 weeks of training, the mean value of the strategies using of the control class and the experimental work was improved, while the average price of the control class was slightly higher than that of the experimental class. The probability was 0.067, more than 0.05, and less than 0.1 by the p-value test. It shows that the strategies using situation in the two classes has improved, which is gradually different, but the significance is not strong [9]. As can be seen from table 3, the frequency of using strategy is low in the control class and experimental work before 12 weeks. After 12 weeks of training, the frequency of using strategies increased. The probability of using the strategy before and after the experiment in the control class and the experimental class was 0 and 0.001 respectively, all lower than 0.05 by the p-value test. It shows that there are differences before and after the experiment, reaching statistical significance. The probability of irrelevance before and after experiment is almost zero. (2) Results of listening and speaking tests In the listening and speaking tests conducted 12 weeks ago and 12 weeks later, the results of the control class and the experimental class were shown in table 4 and table 5 respectively. From the results of statistics of listening and speaking scores in Figure 4, table 4, table 5, the mean value of listening and speaking scores in pretest of the control class and the experimental class were 16.3, 25.03 and 14.74, 25.72 respectively. After p value test, the p value is 0.065 and 0.054, which are between 0.05 and 0.1. It shows that the achievement difference between the two classes is not very significant. After 12 weeks, the mean value of listening and speaking scores in the control class and the experimental class were 21.22, 26.43 and 18.9, 27.03 respectively. The value of p is less than 0.05 by the p value test. This indicates that the listening and speaking effects of the two classes are significantly different, and the listening and speaking effects of the experimental class are more significant [10]. (3) Subjective evaluation of students The significance of changing the attitude of autonomous learning lies in accepting the mode of autonomous learning and consciously entering the state of autonomous learning. It is difficult to achieve the expected learning effect if the change of independent learning attitude only stays in spoken or perfunctory teachers' task requirements. A student from the control class wrote an evaluation of the self-reading task: "medium! The sound quality or the pronunciation of some words is not very good. (Student 1) for the evaluation of the same task, one student from the experimental class wrote, "read the text aloud and read the courseware, so that it is beneficial to the pure pronunciation, correct errors, cultivate the sense of language, and cultivate the ability of self-study." (Student 2) comparing the activities evaluation of two students, we found that the former student was very general. Perhaps he did not know how to improve the level of reading and lacked motivation to insist on reading aloud. Comparatively speaking, the latter student has a deep understanding of reading aloud. In short, the language accumulated through listening and reading can become students' Internalization knowledge and can also become a bridge for them to understand target language. As students become able to learn and learn in the language environment, their chances of leaving teachers for autonomous learning will increase. 5 Conclusion Today, with the rapid development of science and technology and the acceleration of global integration, the demand of society for human resource literacy has changed accordingly. It is a new requirement for the quality of talent in the age of information technology to master and use foreign languages, especially English, to communicate and communicate and to master and use information technology to obtain; process and deal with information. The traditional teaching mode of College English with the goal of acquiring language skills has not been able to meet the needs of the ability of foreign language talents to use foreign languages comprehensively in the development of the times. With students as the center, task as the link and information technology as the means, the goal is to cultivate communicative competence, collaboration ability and improve language skills. The original intention of this research is to build a multi interactive College English teaching mode under the information technology environment and to combine the classroom teaching with autonomous learning. This research takes a university as the research object, and makes an example proving of English listening and speaking. The experimental results show that under the background of information technology, multivariant interactive teaching is conducive to College students' learning and application. 6 Acknowledgement Text of the acknowledgement. This research work is fully supported by School of Foreign Languages in Jinzhong University. 7 References [1] Huffman M K, Schuhmann K, Keller K, et al. Interaction of drift and distinctiveness in L1 English- Informatica 43 (2019) 343–347 347 L2 Japanese learners[J]. Journal of the Acoustical Society of America, 141(5), pp.3517-3517, 2017. https://doi.org/10.1121/1.4987389 [2] Kim S, Jang J, Cho T. Articulatory characteristics of preboundary lengthening in interaction with prominence on tri-syllabic words in American English[J]. Journal of the Acoustical Society of America, 142(4), pp.362, 2017. https://doi.org/10.1121/1.5005132 [3] [3] Huiyong Yang. Comprehensive Evaluation of College English Teaching Mode Based on Online Courses: An Educational Practice from Anhui Polytechnic University[J]. International Journal of Future Generation Communication and Networking, vol. 9, no. 2, pp. 219-230, 2016. https://doi.org/10.14257/ijfgcn.2016.9.2.23 [4] [4] Triantafyllidis A K, Koutkias V G, Chouvarda I, et al. Framework of sensor-based monitoring for pervasive patient care[J]. Healthcare Technology Letters, 3(3), pp.153-158, 2017. https://doi.org/10.1049/htl.2016.0017 [5] [5] Lee Jung Jae; Carson Maggie N; Clarke Charlotte L; Yang Sook Ching; Nam Su Jin.Nursing students’ learning dynamics with clinical information and communication technology: A constructive grounded theory approach.[J]Nurse education today.pp.41-47.2018 https://doi.org/10.1016/j.nedt.2018.11.007 [6] [6 ]Fuentes C D, Dutrénit G. Geographic proximity and university–industry interaction: the case of Mexico[J]. Journal of Technology Transfer, 41(2), pp.329-348, 2016. https://doi.org/10.1007/s10961-014-9364-9 [7] [7] Neto P, Moreira A P. Preface for the special issue on robotics in smart manufacturing[J]. International Journal of Advanced Manufacturing Technology, 85(1-4), pp.1-1, 2016. https://doi.org/10.1007/s00170-014-6028-8 [8] [8] Zhang L, Qin X, Liu P, et al. Estimation of carbon sink fluxes in the Pearl River basin (China) based on a water–rock–gas–organism interaction model[J]. Environmental Earth Sciences, 74(2), pp.945-952, 2015. https://doi.org/10.1007/s12665-014-3788-2 [9] Negash S, Musa P, Vogel D, et al. Healthcare information technology for development: improvements in people’s lives through innovations in the uses of technologies[J]. Information Technology for Development, 24(2), pp.189-197, 2018. https://doi.org/10.1080/02681102.2018.1422477 [10] [10] Armstrong D J, Riemenschneider C K, Giddens L G. The advancement and persistence of women in the information technology profession: An extension of Ahuja's gendered theory of IT career stages[J]. Information Systems Journal, (12), 2018. https://doi.org/10.1111/isj.12185 Modeling the Negotiation of Agents in MAS and Predicting the Performance – an SPE Approach S. Ajitha Ramaiah Institute of Bangalore-54, India E-mail: ajithasankar@gmail.com Keywords: MAS, negotiation, conditional probability, SPE, work load Received: July 15, 2019 Software performance engineering(SPE) process starts at the early stages of software development life cycle which helps to develop software that meets the performance requirements on time and budget. Multi-Agent Systems(MAS) are comprised of one or more agents who coordinate each other to accomplish some task. The coordination can be achieved through cooperation and negotiation. In the early development stages measuring the negotiation workload and predicting the performance remains an important but largely unsolved problem. The problem of uncertainty regarding the negotiation workload is required to be addressed by estimation techniques. Hence, in this research we developed a probabilistic model for the negotiation scenario among the agents in a given time horizon. The negotiation workload obtained from the probabilistic model is integrated with the representative workload of the agents for predicting the performance of agents in MAS. The tool SMTQA is used for obtaining the performance metrics. Analysis of the execution environment is done by considering various configurations in the hardware resources based on the dynamic workload of the negotiation agents over a time horizon. From the sensitivity analysis, the bottleneck resources are identified and suggestions for improvement are proposed. Povzetek: Predstavljena je izvirna metoda za ocenjevanje in napovedovanje delovanja večagentnih sistemov. Introduction A Multi-Agent System (MAS) is usually understood as a system composed of interacting autonomous agents. MAS have been employed successfully in a number of scenarios. The important characteristics of the agents which distinguish it from an object are Autonomous, Cooperation, Goal oriented, Adaptability, Mobility, Negotiation etc. Many articles on MAS have been mainly concerned with functional characteristics such as coordination, rationality and knowledge modeling. The nonfunctional characteristics also have the equal importance as the functional characteristics for any software system [1-5]. This research aims at making a contribution towards the non-functional characteristics Performance of agents in a MAS by considering the negotiation character of agents. Software Performance Engineering (SPE) is a method for constructing software systems to meet the performance objectives at the early stages of Software Development Life Cycle (SDLC). In SPE, system does not exist so it is not possible to develop the work load parameters from measurement data. Therefore, models of the system are used to collect the data required to predict the performance. The different data required for the SPE approach are Workload scenarios, Performance goals, Software Design concepts, Execution environment and Resource usage estimates [5-7]. The performance prediction of agents by considering the cooperation character of agent the authors have published different articles in [8-11]. In the early development stages measuring the negotiation workload and predicting the performance remains an important but largely unsolved problem. The problem of uncertainty regarding the negotiation workload is required to be addressed by estimation techniques. Hence, in this chapter a model for the negotiation scenario among the agents in a given time horizon is developed. The fitness function which represents the fitness of the agent in the negotiation scenario among ‘n’ agents in MAS is considered while framing the model. The negotiation workload obtained from the probabilistic model is integrated with the representative workload of the agents for predicting the performance of agents in MAS [12-14]. The tool SMTQA is used for obtaining the performance metrics. The execution environment is analyzed by considering various configurations in the hardware resources based on the dynamic workload of the negotiation agents over a time horizon. From the sensitivity analysis, the bottleneck resources are identified and suggestions for improvement of the software are made [15]. 2 Methodology A methodology is proposed to model the negotiation scenario of agents in a MAS using probabilistic approach. This is based on the methodology discussed for the distributed systems in [15]. Let t. [t0, T+ t0-1] be the interval, where T be the number of intervals, to be the initial interval of the given time horizon. Let ST be the negotiation services that are considered to be executed during the T intervals. With these assumptions, we have devised the methodology as follows, • Developing a mathematical model for demand of negotiation services over a given time horizon • Modeling the resources in the execution environment • Modeling the variations (alternate designs) in the execution environment • Identifying bottleneck resources and improving the performance by sensitivity analysis 2.1 Modeling of demand of negotiation services Consider a MAS ‘A’ with ‘n’ number of agents. Let ST be a set of negotiation services that to be executed during the T time intervals. Let ST = {S1, S2, S3,……..Sm} be the negotiation services. Let W1, W2, W3,……..Wm be the size (representative workload) of the negotiation services. Let Pijk (t) be the probabilitythat agent ‘i’ communicates with the agent ‘j’ with the work load of Wk at the time interval t . The sum of these Pijk(t) over k equals 1. Each negotiation service that can be occurred in the interval t. [t0, T+ t0-1], is characterized by: • Pijk(t),e – probability of occurrence of the eth negotiation primitive of those specified at time interval t .. • ....,.. – expected demand for each negotiation services s.ST, if the eth primitive occurs among those specified at time interval t. Based on such specification of expected primitives, st demand scenarios can be generated in each interval t, where pij,s(t) be the probability that sth scenario occurs at time t when agent ‘i’ negotiates with the agent ‘j’. In the first period t0: ....0=....0 (1.1) while in the following interval t.[t0, T+ t0-1], the number of scenarios can be recursively computed as: St = Et . St-1 (1.2) Each workload scenario can be defined as the occurrence of one event at period t given one scenario in the previous period t-1. Definition of the demand scenarios based on the specification of six events over a S. Ajitha time horizon constituted of three intervals is given in Table 1. Period Event Probability Demand 0 1 p1 D1 2 p2 D2 1 3 p3 D3 4 p4 D4 2 5 p5 D5 6 p6 D6 Table 1: Definition of the demand scenarios. Figure 1: Conditional probability tree for the workload scenario. 2.2 Calculation of workload The model is simulated by considering three time intervals such that for a given time interval t, t.[0, 2]. The state (negotiation scenario) of the application in time t depends on the state (negotiation service scenario) at the time interval t-1 and the type of the request arrived at t. Hence the scenarios of the negotiation service are considered as states of the software application and the pattern of execution of negotiation services are modeled using the UML, State Chart Diagram. The Figure 2 to Figure 4 represent the workload to be executed during the different time duration. The negotiation services which are having a very less workload are executed during time t=0. During the time t=1 the negotiation services having a higher workload are executed. At time t=2, the negotiation services having an average workload are executed. From Figure 5, it is observed that agent a1 and agent a3 are negotiating more with agent a4. Also agent a2, agent a4, agent a5 are negotiating more with agent a3. Figure 4:Work Load at Time t1. 2.3 Simulation results The scenarios of the negotiation primitives are simulated using the tool SMTQA, and the performance metrics are obtained and tabulated in Table 2. The columns in the table represents average response time (Avg Resp Time), average service time (Avg Serv Time), average waiting time (Avg Wait Time), probability of idle time (Prob Idle) and average dropping of requests (Avg Drop). The rows represent the five agents and the internet. (IntN). From the values it is observed that the number of negotiation services dropped is high in Agent 2 and in the Internet. Hence these two resources are identified as the bottleneck resources. To solve this problem, we conducted the Figure 2: Agents V/S Workloads. Avg Resp Time Avg Serv Time Avg Wait Time Prob Idle Avg Drop Agt 1 0.003 0.003 0.0 0.342 0.0 Agt 2 0.607 0.021 0.586 0.047 0.830 Agt 3 0.006 0.006 0.0 0.196 0.0 Agt 4 0.008 0.007 0.001 0.181 0.004 Agt 5 0.016 0.013 0.002 0.218 0.005 IntN 0.060 0.021 0.586 0.047 0.830 Table 2: Simulation Result for the configuration C1. sensitivity analysis by considering different configurations and are presented in the next section. 2.4 Sensitivity analysis Simulation of the behavior of the resources is carried out by considering the configurations C1 to C6 is as follows. C1:-Processing speed of CPU is 2000, and the Internet speed assumed is 96. C2:-Processing speed of CPU is 3000, and the Internet speed assumed is 96. C3:-Processing speed of CPU is 4000, and the Internet speed assumed is 96. C4:-Processing speed of CPU is 2000, and the Internet speed assumed is 146. C5:-Processing speed of CPU is 3000, and the Internet speed assumed is 146. C6:-Processing speed of CPU is 4000, and the Internet speed assumed is 146. The results of the different simulation runs are presented in the form of tables. The results obtained for Agent 1 for the different configurations considered is presented in Table 3. The maximum time taken by the Agent 1 to respond is 0.036 in the configuration C4 and minimum time taken to respond is 0.003 with configuration C1. The waiting time in Agent1 is also maximum for the configuration C4. This is due to the configuration of C1; the number of negotiation services dropped is more due to the low configuration of the Internet. Hence the number of negotiation services that are processed by Agent 1 is less compared to other configurations. In configuration C4, the processing speed of the Internet is increased so that the negotiation services by Agent 1 are more. Agent 1 Avg Res Time Avg Serv Time Avg Wait Time Pro Idle Avg drop C1 0.003 0.003 0 0.342 0 C2 0.02 0.019 0.001 0.196 0 C3 0.013 0.012 0 0.132 0 C4 0.036 0.032 0.004 0.135 0 C5 0.017 0.017 0 0.271 0 C6 0.016 0.015 0.001 0.137 0.003 Table 3: Simulation results obtained forAgent 1 with different Configuration. The results obtained for Agent 2 for the configurations considered are tabulated in the Table 4. Agent 2 has taken the maximum time 0.607 to respond under Configuration C1 and the minimum time to respond is 0.009 with configuration C5. Figure 21 presents the average dropping of requests and probability of idle time of Agent 2. The maximum number of requests is dropped in configuration C1. The maximum waiting time for the requests is observed with configuration C1. This has happened because Agent2 has received more requests. Agent 2 Avg Res Time Avg Serv Time Avg Wait Time Pro Idle Avg drop C1 0.607 0.021 0.586 0.047 0.83 C2 0.02 0.15 0.004 0.108 0.005 C3 0.01 0.009 0.001 0.123 0 C4 0.023 0.018 0.005 0.115 0.005 C5 0.009 0.008 0.001 0.159 0 C6 0.01 0.009 0.001 0.133 0 Table 4: Simulation results obtained forAgent 2 with different Configuration The results obtained for Agent 3 for the configurations considered are presented in Table 9.5. Agent 3 has taken the maximum time 0.118 to respond under Configuration C4 and the minimum time to respond is 0.003 with configuration C5. The average dropping of requests and probability of idle time of the Agent 3 is plotted in Figure 23. The maximum number of requests is dropped in configuration C4. The maximum waiting time for the requests is observed in configuration C4 for Agent 3. Agent 3 Avg Res Time Avg Serv Time Avg Wait Time Pro Idle Avg drop C1 0.006 0.006 0 0.196 0 C2 0.008 0.008 0.001 0.14 0 C3 0.094 0.052 0.042 0.083 0.073 C4 0.118 0.06 0.059 0.082 0.102 C5 0.003 0.003 0 0.255 0 C6 0.055 0.033 0.022 0.099 0.044 Table 5 Simulation results obtained forAgent 3 with different Configuration. The Table 6 presents the results obtained for Agent 4 for the different configurations considered Agent 4 has taken the maximum time 0.033 to respond under Configuration C3 and the minimum time to respond is 0.008 with configuration C1. The maximum number of requests is dropped in the configuration C3. The maximum waiting time for the requests is observed with configuration C3. Agent 4 Avg Res Time Avg Serv Time Avg Wait Time Pro Idle Avg drop C1 0.008 0.007 0.001 0.181 0.004 C2 0.018 0.016 0.002 0.125 0 C3 0.033 0.025 0.008 0.116 0.011 C4 0.032 0.025 0.007 0.116 0.009 C5 0.014 0.013 0.001 0.179 0.004 C6 0.015 0.013 0.002 0.118 0 Table 6: Simulation results obtained forAgent 4 with different Configuration. The Table 7 presents the results obtained for Agent 5 for the different configurations considered. Agent 5 has taken the maximum time 0.07 to respond under Configuration C4 and the minimum time to respond is 0.016 with configuration C1 and C5. The maximum number of requests is dropped in the configuration C4. The maximum waiting time for the requests is observed with configuration C4. It is observed that the response times of Agent 3, Agent 4 and Agent 5 in the considered configuration are closer to each other and only a few numbers of negotiation services are dropped. Maximum response time is experienced with Configuration C3 and C4 and observed that the dropping of requests in the Internet is the least for the configurationC3 and C4. Lowest Response time is in C1 for Agents A1, A3, A4, A5 but experienced highest number of dropping in the requests. The behavior of Agent 2 is observed different compared to all the other agents’ behavior. The reason can be that the average workload of Agent 2 is less compared to the workload of other agents, because of that Agent 2 could execute all the negotiation Agent 5 Avg Res Time Avg Serv Time Avg Wait Time Pro Idle Avg drop C1 0.016 0.013 0.002 0.218 0.005 C2 0.022 0.018 0.004 0.211 0.004 C3 0.025 0.02 0.005 0.107 0.002 C4 0.07 0.045 0.026 0.091 0.03 C5 0.016 0.013 0.003 0.128 0.008 C6 0.025 0.021 0.004 0.126 0 Table 7: Simulation results obtained forAgent 5 with different Configuration. INTERNET Avg Res Time Avg Serv Time Avg Wait Time Pro Idle Avg drop C1 0.0607 0.021 0.586 0.047 0.83 C2 0.088 0.001 0.087 0.051 0.413 C3 0.025 0.021 0.004 0.033 0.068 C4 0.07 0.004 0.066 0.048 0.22 C5 0.163 0.004 0.159 0.05 0.674 C6 0.068 0.004 0.064 0.048 0.198 Table 8: Simulation results obtained for Internet with different Configuration. requests it received. Also we observed that when the Internet speed is increased the dropping of requests reduced which gives the inference that many negotiation requests are executed by the agents successfully. 3 Summary In this work, we presented a methodology to model the negotiation between the agents and predicting the performance of the system. We presented methodology to: i) develop a mathematical model for the workload of negotiation scenarios over a time horizon, ii) modeling the execution environment, iii) and iv) analyzing the execution environment for variations in resource configurations. The sensitivity analysis is done by considering modification in the resource configuration one at a time, and it also describes bottleneck resources. The output showed that how the different configurations of resources affect the response time of the agents. 4 References [1] A.Dorri, S. S. Kanhere and R. Jurdak, "Multi-Agent Systems: A Survey," in IEEE Access, vol. 6,pp.28573-28593,2018. https://doi.org/10.1109/ACCESS.2018.2831228 [2] Jing Xie & Chen-Ching Liu (2017) Multi-agent systems and their applications, Journal of International Council on Electrical Engineering, 7:1,188-97,DOI:10.1080/22348972.2017.1348890 https://doi.org/10.1080/22348972.2017.1348890 [3] Wooldridge, M.: An Introduction to Multi-Agent Systems. John Wiley Sons, Inc. New York, NY, USA (2001) [4] Ebrahim AlHashel, "A Conceptual agent Cooperation Model for Multi-agent Systems' Team Formation Process", Third 2008 International Conference on Convergence and Hybrid Information Technology, pp. 12-20,2008. https://doi.org/10.1109/ICCIT.2008.367 [5] Connie U.Smith and Lioyd G. Williams., "Building Responsive and Scalable Web Applications." December 2000. Proceedings CMGC. [6] S. Balsamo, A. D. Marco and P. Inverardi., "Model-Based Performance Prediction in Software Development: A Survey." IEEE Transactions on Software Engineering, May 2004, Vols. Vol. 30, No.5. https://doi.org/10.1109/TSE.2004.9 [7] V. Cortellessa and R.Mirandola., "Deriving a Queueing Network Based Performance Model from UML Diagrams." s.l. : ACM Proc. intl, 2000. Workshop Software and Performance. pp. pp. 58-70. https://doi.org/10.1145/350391.350406 [8] Ajitha S, Suresh Kumar T.V, Rajanikanth K. A Quantitative Framework for early prediction of Cooperation in Multi-Agent System. ICTACT Journal on Soft Computing 2013; 587-595, DOI: 10.21917/ijsc.2013.0085. https://doi.org/10.21917/ijsc.2013.0085 [9] S. Ajitha, Dr.T.V.Suresh Kumar, Dr.K.Rajanikanth, "Artificial Neural Network Approach for predicting performance of MAS using SPE approach". International Journal of Software Engineering, Volume6, No.2 ,pages 3-20, July 2013. [10] S. Ajitha, Dr.T.V.Suresh Kumar, D.E.Geetha, Dr.K.Rajanikanth "Modeling Co-operative Index of Multi-Agent Systems using Execution Graph". Proceedings of International Conference on Advances computing in Intelligent Systems and Computing Volume 174,2012, pp41-48, Springer,DOI:10.1007/978-81-322-0740-5. https://doi.org/10.1007/978-81-322-0740-5 [11] S. Ajitha, Dr.T.V.Suresh Kumar, D.E.Geetha, Dr.K.Rajanikanth "Early Performance Prediction of Co-operative Multi-Agent Systems" procedia Engineering,38(2012)3037-3048,DOI:10.1016/ j.proeng.2012.06.354. https://doi.org/10.1016/j.proeng.2012.06.354 [12] Ye Chen, Yun Peng, Tim Finin, Yannis Labrou, Bill Chu, Jian Yao, Rongming Sun, BobWillhelm, Scott Cost, A negotiation-based Multi-agent System for Supply Chain Management ,In Proceedings of Agents 99 Workshop on Agent Based Decision-Support for Managing the Internet-Enabled Supply-Chain. [13] T. Wong, C. Leung, K. Mak, and R. Fung, "An agent-based negotiation approach to integrate process planning and scheduling," International Journal of Production Research, vol. 44, no. 7, pp. 1331-1351,2006. https://doi.org/10.1080/00207540500409723 [14] W L Yeung, Performance of Time-Bound Negotiation in Agent-Based Manufacturing Control, Proceedings of the World Congress on Engineering 2012 Vol III WCE 2012, July 4 -6, 2012, London, U.K. [15] D.E Geetha, T.V. Suresh Kumar, Performance Modeling and evaluation of Distributed Systems, Ph.D thesis, Visvesvaraiah Technological University, Karnataka, 2012 A Novel Agent Based Load Balancing Model for Maximizing Resource Utilization in Grid Computing Ali Wided and Kazar Okba Department of Computer Science, Mohamed Khider University, Biskra, Algeria E-mail: aliwided1984@gmail.com Keywords: grid computing, load balancing, multi agent system, performance metrics, agent based load balancing Received: July 15, 2019 Grid is the collection of geographically distributed computing resources. For effective management of these resources, the manager must maximize its utilization, which can be achieved by efficient load balancing algorithm, The objective of load balancing algorithms is to assign the load on resources to optimize resource use while reducing total jobs execution time. The proposed agent based load balancing model aims to take advantage of the agent characteristics to generate an autonomous system. It also addresses similar systems drawbacks such as instability, scalability or adaptability. The performance of the proposed algorithms were tested in Alea 2 simulator by using different parameters such as response time, resources utilization and overall queue time. The performance evaluation suggests that the proposed algorithm can enhance the overall performance of grid computing. Povzetek: Predstavljena in s simulatorjem analizirana je agentna metoda razporejanja obremenitev v omrežju. Introduction Due to the emergence of grid computing on the Internet, a hybrid load balancing algorithm, which takes into account various factors such as grid architecture, computer heterogeneity, communication delays, network bandwidth, resource availability, unpredictability and job characteristics, is now required. For grids, scalability and adaptability are two major issues. As for the centralized resource scheduling problem, the limitation of scalability and computational performance is inevitable. Moreover, due to resource heterogeneity, resource variations, application diversity and grid environments are dynamic. Therefore, adaptive and robust scheduling techniques are preferred [1][2]. Multi-agent systems offer promising features for resource managers. The reactivity, proactivity, scalability, cooperation, robustness, flexibility and autonomy that characterize agents can help in the complex task of managing resources in dynamic and changing environments. This paper presents a new Agent Based Load Balancing Algorithm, called ABLBA. A hierarchical architecture with coordination is designed to ensure scalability and efficiency. In addition, a multi-agent approach is applied to improve the adaptability. The proposed algorithm aims to reduce the average response time, as much as possible, of jobs submitted to the Grid, and to maximize throughput and resource utilization. Related works Authors in [3] proposed a multi-agent load balancing model by analyzing the load of compute nodes and the subsequent migration of virtual machines from overloaded nodes to underloaded nodes. The proposed system involves multiple nodes that interact to implement MapReduce jobs. The multi-agent system consists of a group of agents: node sensor agent, simulation model sensor agent, analysis agent, migration agent and distribution agent. Analysis and distribution agents are defined as reasoning agents. In [4], a decentralized computing algorithm was proposed to assign and schedule jobs on a distributed grid. Using the properties of multi-agent systems, the proposed distributed resource allocation protocol (dRAP) is described as follows: An agent in the system is simply a node. Each agent has a vector including the number of CPUs in its cluster and the residual time to complete the execution of its current process. Each agent is assured to be in exactly 1 out of 4 cases during the simulation. A main feature of this algorithm is that nodes ask their neighbors to form clusters. This reduces waiting time and communication costs. One optimization to consider would be to delay the disconnection of the cluster in state 4, which would guide learning or memory in the system where the planner would be able to remember the requirements of the past process. The problem with this algorithm is its decentralized nature, it is neither a centralized control nor a precise synchronization on nodes (agents). The study in [5] presented the development of an agent-based model for managing network resources with defined operations so that the user can perform jobs efficiently and effectively and thus significantly improve management by a gLite Grid middleware. The proposed solution provides a platform based on a collection of agents in a virtual organization. The key aspects of this proposal architecture are: resource tracking, load balancing and agent hierarchy. In [6] the authors proposed a new load balancing structure based on the moving agent and a technique for optimizing ant colonies. In the proposed structure, a dispatcher agent is involved in distributing the tasks received to the worker agents according to the right decisions to minimize the overall execution time (makespan). The proposed framework is constructed using three layers which are the producer of user tasks, the scheduling load balancing layer and the workers' layer. This study should be complemented by comparing their results with other methods, minimizing task movements and resulting in additional costs in the migration process. Authors in [7] presented the design and implementation of a priority scheduling and fuzzy load balancing model in a computing grid. In this grid template, the user sends his jobs to the grid agent, after the grid scheduler uses the priority-based scheduling algorithm to schedule jobs from the grid agent to the available resource. Load balancing is done using the fuzzy logic technique Propose, in which a set of fuzzy rules are produced using the resource and the work parameter. As fuzzy control rules are collected using linguistic variables, perceptual knowledge and inspection are easily integrated into the control mechanism. 3 Proposed agent based load balancing model A grid computing was modelled as a set of clusters. Each cluster was composed of nodes and belonged to a LAN local domain (Local Area Network). Every cluster was connected to the WAN global network (World Area Network) by a Switch [8]. The proposed Agent Based load balancing model was based on mapping the Grid architecture into a tree structure. This tree was built by aggreGAtion as follows: first, for each cluster, a two level subtree was created. The leaves of this sub-tree correspond to the cluster nodes, and its root, called cluster manager, represents a virtual node associated with the cluster. Secondly, sub-trees corresponding to all clusters were collected to generate a three level sub-tree whose root is a virtual node designated as a Grid manager. The concluding tree is referred to as C/N, where C is the number of clusters that constitute the Grid and N the number of worker nodes [8]. This study aims to develop a hierarchical load balancing model based on a multi-agent system. There are two key challenges for Grid computing: heterogeneity and scalability. The authors propose a three-layer architecture to address the scalability issue. Connecting or disconnecting resources (worker nodes or clusters) correspond to simple operations in a tree (adding or removing leaves or sub-trees). The proposed agent based load balancing model aims to take advantage of the agent’s characteristics to create an autonomous system. It also addresses similar A. Wided et al. disadvantages such as instability, scalability, adaptability, etc., and other specific issues related to grid computing. 3.1 Model characteristic The proposed model is characterized as hierarchical; this characteristic facilitates the circulation of information through the tree and defines the flow of messages in the proposed strategy. Three types of load information movements can be identified: • Ascending movement: this movement relates to the load information movement, to get current load state. from Level 2 (node Agents) towards Level 1 (Cluster Agents). or from Level 1(Cluster Agents) towards Level 0 (grid Agents). With this movement, the cluster manager can have a global view of the cluster load or the grid manager can have a glob view of the grid load. • Horizontal movement: it concerns the useful parameters for the execution of load balancing operations. This movement relates to task assignment intra-cluster in Level 2. • Descending movement: this movement allows to take decisions for task assignment or jobs migration, the decisions taken by cluster Agents at levels 1 to the Migration Agents at same level. And from Migration Agents at level 1 to Node Agents at level 2, also from Grid Agent at level 0 to Cluster Agents in level 1. The proposed model: • supports the scalability and heterogeneity of grids: insertion or elimination entities (processing elements, nodes or clusters) are very simple operations in the proposed model (insertion or elimination nodes, subtrees); • is totally independent of any physical structure of a grid: the conversion of a grid into a tree is a unique conversion. Each grid corresponds to one and only one tree; • is based on the exchange of information between Nodes and clusters through their respective agents. Level 0: At this level, Grid Agent is located, the Grid users send their jobs to the Grid Agent, for which it is responsible: • receiving jobs from Grid users • sending jobs for Node Agents • all Cluster Agents are started by Grid Agent • initiating a global load balancing process Level 1: At this level, Cluster Agent is associated with a physical grid cluster; this Agent is responsible for: • the maintenance of the load information relating to each of its Node Agents. • estimating the load of the associated cluster and sending this information to Grid Agent. • the decision to start local load balancing • sending load balancing decisions to Migration Agent • Migration Agent is started by its associated cluster Agent • all Node Agents are started by their corresponding Cluster Agent Migration Agent is also present at this level, whose role is to: • start the migration process • send the migration decisions to the Node Agents. • wait for an acknowledgement from receiver node and ensure that the migrated jobs are received and successfully resumed at the destination node Level 2: At this level, Node Agent is present; it is necessary to have one Node Agent on each node; every Node Agent at this level is responsible for: • maintaining its load information • sending this information to its associated • Cluster Agent • working in cooperation with the Migration • Agent to execute the migration process • collect information about the jobs (number of jobs queued at node, arrival time, waiting time, submission time, start time, processing time and finish time of each job on the local node) • remove the terminated, leaving or migrated jobs from queue of jobs • calculate the total load of node • receive jobs sent by Grid Agent 3.2 Proposed algorithms According to the proposed model, two levels of load balancing are considered: Intra-cluster Agent based load balancing algorithm and Inter-Clusters Agent based load balancing algorithm. There are certain specific events that change the load configuration in Grid computing and can be classified as follows: • Any new job is arrived • Accomplishment of execution of any job • Any new node is arrived • Any existing node is removed • Failure of Machine at any node • The node become overloaded When any of these events happen, the local load value is changed. Table 1 summarizes the notations used in the proposed algorithms. Parameter Description N Node LoadN Load of Node Qlength Queue length CPU-U CPU utilization of Node Mem Memory utilization of node THH The higher threshold THL The lower threshold OLD-list Overloaded List ULD-list Underloaded List BLD-list Balanced List Loadavg Average Load NBRN Number of Nodes of cluster C Cluster Table 1: Notations used in the proposed algorithms. 3.2.1 Intra-cluster agent based load balancing algorithm Depending on its current load, each Cluster Agent decides to start a Job Migration operation. In this case, the Cluster Agent tries, in priority, to balance its load among its nodes. Load estimation The node load at a given time was simply described by the CPU queue length. It indicates the number of processes awaiting execution. The proposed algorithm considers CPU-U (CPU Utilization), Q length (Queue length) and Mem (memory utilization) as load information parameters to measure the load of a node. These parameters are calculated as follows: Load (CPU-U)= (U1+U2+……+UT)/T, where: U1+U2+……+UT is the value of CPU-U in a previous one second interval. Load (Qlength) = (Q1+Q2+…...+QT)/T, where: Q1,Q2,……...,QT is the value of Qlength in a previous one second interval. Load(Mem)=(M1+M2+……...+MT)/T Where: M1,M2,……...,MT is the value of Mem in a previous one second interval. T is the number of time intervals. The averaged information of CPU-U, Qlength and Mem are the load parameters used to describe the node load. Algorithm 1. An algorithm for Node Agent 1: T‹5 seconds 2: Waiting for jobs; 3: Create jobs queue for related node; 4: In each one second of T intervals do 5: Calculate (CPU-U); 6: Calculate (Qlength); 7: Calculate (Mem); 8: End do 9: Load (CPU-U) = (U0+U1+…..UT)/T; 10: Load (Qlength) = (Q0+Q1+…..QT)/T; 11: Load (Mem) = (M0+M1+…..MT)/T; 12: Send load information for related Cluster Agent 13: Wait for load change // happening of any of defined events 14: If (events_happens ()=1 or events_happens ()=4) then // Termination or migration of job 15: Remove terminated or migrated job from the waiting queue 16: Subtract their load value from the total local load of node. 17: Send new load to its Cluster Agent associated; 18: End if 19: If (events_happens ()=2 or events_happens ()=3) then // new or incoming job 20: Add the newly created or incoming job for the waiting queue 21: Add their load value for the total local load of node 22: Send new load to its Cluster Agent associated; 23: End if Function events_happens () output Type: integer 1: If (Job.state=Termination) then events_happens () =1; End If 2: If (Job.state=Start) then events_happens () =2; End If 3:If (Job.state=Incoming Migrating) then events_happens ()=3; End If 4: If (Job.state = migrated) then events_happens ()=4; End If 5:If (Arrival of any new resource) then events_happens ()=5; End If 6: If (Cluster.state=saturated )then events_happens ()=8; End If 7:If (Cluster.state=unbalanced) then events_happens ()=9; End If Location policy In the next step, the nodes must be classified according to their load. Three states were used for classification: overloaded, underloaded and balanced. First, Cluster A. Wided et al. Agent must calculate two threshold values, which are calculated as follows: • cluster Agent calculates load average of each parameter (CPU-U and Qlength) over all related nodes. • Loadavg(Qlength)=(load1+load2+….loadNBRN)/NBR N, where Loadavg(Qlength) is the average load of Qlength over all related nodes. • load1,load2,….loadn are the current Qlength of each node calculated by Node Agent. • Loadavg (CPU-U) =(load1+load2+….loadNBRN)/ NBRN, where Loadavg (CPU-U) is the average load of CPU-U over all related nodes. • load1,load2,….loadNBRN are the current load of CPU-U of each node calculated by Node Agent. Calculation of threshold values The higher and lower threshold values of Qlength and CPU-U of parameters are calculated by multiplying the average load of (Qlength or CPU-U ) and a constant value. • THH(Qlength) =H*Loadavg(Qlength) • THL(Qlength) =L* Loadavg(Qlength) • THH(CPU-U) =H*Loadavg(CPU-U) • THL(CPU-U) =L* Loadavg(CPU-U) where, THH is the high threshold and THL is the low threshold. H and L are constants. The next step is to divide the nodes for balanced, overloaded and underloaded nodes using the threshold values as follows: • Overloaded: the node will be added for overloaded list if queue length is high, or CPU utilization is high, or memory usage is greater than 85%, then the node is classified as overloaded node. • Underloaded: the node will be added for underloaded list if queue length is low, or CPU utilization is low. • Balanced: the node is not into the overloaded list or the underloaded list. The node is in a balanced load state. They are considered to be more loaded than the low state and less loaded than the high state. Algorithm 2. An algorithm for Cluster Agent 1: Startup its related Node Agent 2: Startup its related Migration Agent 3: Receive load information(LoadN(Qlength), LoadN(CPU-U)) from its related nodes. 4: Calculate and send its load information for Grid Agent. 5: somme ‹0; somme1‹0; 6: For every Node N of cluster C do 7: Somme‹ Somme+ LoadN(Qlength); 8: Somme1‹ Somme1+ LoadN(CPU-U); 9: End For 10: Loadavg(Qlength)= somme1/NBR-N; 11: Loadavg(CPU-U)= somme/NBR-N; 12: THH(Qlength)= Loadavg(Qlength)*H; 13: THL(Qlength)= Loadavg(Qlength)*L; 14: THH(CPU-U)= Loadavg(CPU-U)*H; 15: THL(CPU-U)= Loadavg(CPU-U)*L; 16: Partition Nodes into overloaded list OLD­ list, underloaded list ULD-list and balanced list BLD-list 17: OLD-list‹Ř; ULD-list‹Ř; BLD-list‹Ř; 18: For every Node N of cluster C do 19: If ((LoadN(Qlength)>THH(Qlength)) or (LoadN(CPU-U)>THH(CPU-U))or (Load(Mem)>85%)) then 20: OLD-list ‹OLD-list .N; 21: End If 22: Else If ((LoadN(Qlength))< THL(Qlength))or(LoadN(CPU­U)