https://doi.org/10.31449/inf.v48i16.6371                                             Informatica 48 (2024) 13–26   13 
Automated Logistics Control Model Based on Improved Ant Colony 
Algorithm 
 
Shijie Ye, Na Liu* 
School of Business, Chongqing City Management College, Chongqing 401331, China 
E-mail: liuna_cswu@yeah.net  
*
Corresponding author 
Keywords: ant colony algorithm, automated logistics, logistic chaos mapping, path optimization, stacker crane 
Received: June 5, 2024 
With the rapid development of modern logistics industry, traditional automated logistics control systems 
often lack transparency and visualization of the entire supply chain. It cannot comprehensively manage 
and optimize the entire supply chain. Therefore, an automated logistics control model is constructed based 
on ant colony algorithm and logistic chaotic mapping. By simulating the pheromone transmission process 
of ants, the optimal logistics transportation path is found. From the experimental results, the improved 
Ant Colony Algorithm (ACA) was tested on the DT100dataset, achieving an optimal solution within 200 
iterations. Compared with traditional methods, the cost was reduced by 0.25 units. The distance solution 
image of the improved ACA was overall concave downwards, with a significant decrease after 20 
iterations. The designed automated logistics control system used map APIs and sensor data to obtain 
real-time delivery routes. On average, each logistics node consumed 1.01% of electricity. Compared with 
traditional methods, it had the highest prediction accuracy, with an R2 of 0.98. In summary, the improved 
ant colony algorithm can optimize logistics delivery paths, reduce delivery time and cost, improve 
delivery efficiency, and reduce delivery energy consumption. 
Povzetek: Razvit je avtomatiziran logistični nadzorni model za optimizacijo logističnih poti, ki zniža 
stroške ter izboljša dostavo. Temelji na izboljšanem algoritmu mravelj in kaotičnem logističnem 
preslikavanju.
1 Introduction 
With the development of the global economy, the 
application of logistics industry in various fields is 
becoming increasingly widespread [1]. Traditional 
logistics control methods often exhibit certain limitations 
when dealing with complex and ever-changing 
environments [2]. In this context, improving the 
efficiency and accuracy of logistics control to adapt to 
complex and changing environments has become a 
research focus. Traditional automated logistics control 
systems mainly rely on precise sensors, actuators, and 
complex control algorithms. However, these systems 
often have some shortcomings in dealing with the 
uncertainty and complexity of actual production 
processes. For example, traditional control algorithms do 
not achieve ideal results in dealing with complex and 
nonlinear logistics problems. Logistic chaotic mapping 
can generate stochastic sequences. In addition, logistic 
chaotic mapping can also be used to optimize logistics 
path planning, improving the efficiency and accuracy of 
logistics transportation [3]. Ant Colony Algorithm (ACA) 
can quickly find the optimal solution by introducing some 
new rules. Each ant can prioritize the path with higher 
concentration of pheromones. Therefore, this algorithm 
has strong global search ability [4-5]. For this purpose, 
the study combines logistic chaotic mapping with 
improved ACA to optimize automated logistics control. 
An automated logistics control model is proposed. This 
study explores how to apply the improved ACA to 
optimize logistics transportation by analyzing its 
principles and application methods. The innovation of 
this study combines logistic chaos mapping and improved 
ACA, utilizing the efficiency and accuracy advantages of 
automated logistics control model operation. Ultimately, 
it achieved device automation and intelligent control, 
while reducing system cost and maintenance difficulties. 
The study involves four parts. The first part is to analyze 
the current research status of improved ACA. The second 
part describes the application of improved ACA in 
automated logistics control and the process of model 
construction. The third part analyzes the performance of 
the improved ACA and the designed logistics model. The 
last part summarizes the entire text. 
2 Related works 
To solve the installation cost and environmental issues of 
renewable energy, Güven et al. introduced a powerful 
rule-based energy management scheme based on ACA to 
manage the power flow between the system components 
that make up the micro-grid. Experiments showed that 
this method had good performance and convergence, with 

14   Informatica 48 (2024) 13–26                                                                    S. Ye
 
et al. 
lower system cost [6]. To improve transient stability 
using continuous domain ant colony optimization 
algorithm, Moradi et al. solved the optimal power flow 
problem with transient stability constraints and extracted 
sensitivity coefficients. Compared with other traditional 
methods, the research method reduced the fuel cost of 
power system operation to 60928.36$/h [7]. To solve the 
communication delay problem between controllers and 
switches caused by link failures in the network, Li et al. 
established a model based on task delay and dynamic 
constraints. A heuristic ACA was used to solve the 
adaptive allocation scheme of computing resources. 
Experimental showed that the proposed method 
optimized computational resources [8]. To solve the 
manual scheduling difficulties and delivery delays in the 
production process, He et al. proposed an improved ACA 
to establish a weaving production scheduling model. 
Experimental showed that this method optimized the 
actual weaving scheduling [9]. Pontes et al. proposed a 
new linear discriminant analysis variable selection 
algorithm based on ACA to increase the probability of 
discarding non informative variables. Experiments 
showed that this method could minimize generalization 
problems related to multi-collinearity to the greatest 
extent possible [10]. 
The application of green logistics in the offshore oil 
logistics industry is becoming increasingly urgent. Zhu et 
al. conducted empirical analysis from the basic concept of 
green logistics. Then a series of measures for the 
development of green logistics in the offshore oil 
logistics industry were proposed [11]. To ensure the 
sustainability of drug distribution, Jin et al. proposed a 
hybrid replenishment model based on the multi-step 
traveling salesman problem model and the vehicle routing 
problem model. The results showed that the backfill route 
length under this model reduced by 32.18% [12]. To 
solve the problems that arise in the construction of 
logistics networks, Zhang et al. defined the network level, 
node functions, and connections between nodes based on 
the logistics network structure. Finally, the cost 
optimization in the logistics network was achieved 
through an improved urban agglomeration order gravity 
model [13]. Defalque et al. established a multi-objective, 
multi product, multi-level, and multi cycle mixed integer 
linear programming model to optimize the waste paper 
logistics process in the intermediate center. The model 
improved the efficiency of sustainable logistics processes 
[14]. Considering the time dependent effects caused by 
traffic congestion in logistics, Guo et al. established a 
time window green vehicle routing problem for cold 
chain logistics based on adaptive large neighborhood 
search technology. The research algorithm was effective 
and superior [15]. The summary table of the above 
research results is shown in Table 1. 
 
 
Table 1: Summary table of research results 
Study Performance Boundedness 
G üven et al. [6] 
Compared with the three methods of 
HOMER, ant colony optimizer, and Jaya, it 
has better performance and convergence 
The complexity of algorithms is increased, 
resulting in lower computational efficiency 
Moradi et al. [7] 
The fuel cost for power system operation is 
reduced to 60928.36 $/h 
It requires a significant amount of 
computing resources and memory space, 
with high requirements for computer 
hardware 
Li et al. [8] 
It can optimize computing resources and 
reduce task completion delays 
Its ability to handle complex nonlinear 
problems is limited 
He et al. [9] 
It can optimize the actual weaving 
scheduling 
The complexity of algorithms has 
increased, resulting in lower computational 
efficiency 
Pontes et al. [10] 
The classification accuracy in datasets 
involving wide absorption bands in the 
ultraviolet region, low resolution, and 
strong spectral overlap is over 90% 
It is sensitive to initial parameter settings 
and adjustment process 
Zhu et al. [11] 
It provides countermeasures and measures 
for the development of green logistics in 
the offshore oil logistics industry 
There is no factual evidence to prove the 
effectiveness of the proposed measure 
Jin et al. [12] 
The length of the backfill route is reduced 
by 32.18% 
Algorithm complexity increases and 
computational efficiency decreases 
Zhang et al. [13] 
It aims to optimize costs and achieve 
linkage of logistics network structure 
There is no factual evidence to prove the 
effectiveness of the proposed method 
Defalque et al. [14] 
Effective solutions are proposed in the 
collected instance calculation tests 
This study lacks comparative experiments 
Guo et al. [15] 
The proposed solution is 14.14% better 
than the CPLEX solution, with a shorter 
Algorithm complexity increases 

Automated Logistics Control Model Based on Improved Ant… Informatica 48 (2024) 13–26 15 
solution time 
 
 
In summary, ACA has high sensitivity, which is widely 
used in machine automation collaboration. The current 
logistics control system needs further improvement in 
automation and intelligence. This study aims to improve 
the ACA and apply it to automated logistics control 
systems to improve logistics efficiency and reduce 
operating cost. 
 
3 Application of improved ant colony 
algorithm in automated logistics 
control model 
Traditional logistics control models are difficult to handle 
rich customer and order information data. Therefore, 
ACA is introduced into the logistics control model. Some 
improvement measures are taken to address the 
shortcomings of traditional ACA, which effectively solve 
problems such as path planning and resource scheduling 
in the logistics process. 
 
3.1 Ant colony algorithm optimization 
strategy for automated logistics control 
With the increasingly fierce market competition, various 
links in the production process are becoming more 
specialized [16]. This trend has led to the separation of 
logistics and commercial flow, gradually highlighting the 
importance of logistics. Traditional logistics control 
models lack automation and intelligent support. Many 
operations and management require manual labor, which 
is inefficient and prone to errors. At the same time, the 
lack of intelligent support also makes it difficult for 
enterprises to predict and make decisions on logistics 
processes. This also prevents timely adjustment and 
optimization of logistics strategies [17]. On this basis, the 
automated logistics control model has emerged. The 
automated logistics control model is a comprehensive 
system that utilizes various automation technologies and 
equipment to control and optimize the logistics process. 
As an important component of logistics chain cost, 
managing and controlling logistics transportation cost can 
optimize logistics processes and improve logistics 
transportation efficiency. The logistics transportation cost 
is shown in equation (1). 
1
11
km
ki
ki
C S f
==
=

       (1) 
In equation (1), 
k
S is the variable. k is the vehicle 
number. 
m
 represents the number of items that includes 
all transportation costs. f is the driving cost. The 
corresponding transportation cost is shown in equation 
(2). 
2
1 0 1
n n k
i ijk ijk
i j k
C C X
= = =
=
  
   (2) 
In equation (2), 
ijk
C
 represents the transportation cost 
between points i and 
j
. 
ijk
C
 is the variable between 
transportation points. 
n
 is a constant. The automated 
logistics control system is shown in Figure 1. 
 
Front desk work 1 Front desk work 2 Front desk work n Robot control
Goods warehouse Cubic warehouse management
PLC monitoring
PLC monitoring
AVG monitoring
Sorting system
 
Figure 1: Automated logistics control system 
 
The automatic logistics control system includes sorting 
system, goods warehouse, PLC monitoring, 
three-dimensional warehouse management, AGV 
monitoring, server, front-end work, mechanical arm, and 
other parts. The sorting system is mainly used to identify 
and classify goods. According to their destination or type, 
they are distributed to different areas or modes of 
transportation. Goods warehouse is a facility used for 
storing goods, including automated warehouses and 
traditional warehouses. Logistics system monitoring 
mainly includes PLC monitoring and AGV monitoring. 
PLC monitoring can be used as a programmable logic 

16   Informatica 48 (2024) 13–26                                                                    S. Ye
 
et al. 
controller to monitor and control various parts of the 
logistics system. AGV monitoring is used to transport 
goods within or between warehouses, which is 
responsible for monitoring its operational status and 
position. The entire logistics system implements 
three-dimensional warehouse management, tracking and 
managing goods through computer systems to ensure 
accurate storage and retrieval of goods. The front-end 
work mainly includes processing orders, querying 
inventory, etc. To perform various tasks such as grabbing, 
moving, assembling, etc., the system uses a robotic arm 
for operation. With the increasing demand for logistics 
technology, the management level of traditional 
automatic logistics systems is relatively low. The ability 
to control the supply chain, plan transportation networks, 
and optimize warehouse management needs to be further 
improved. ACA is an optimization algorithm that 
simulates the process of ants searching for food in nature. 
It achieves information sharing and collaboration during 
the optimization process by simulating the pheromone 
transmission process of ants [18-19]. Therefore, 
introducing ACA into the logistics control model can 
effectively solve problems such as path planning and 
resource scheduling in the logistics process, improving 
the management level of automatic logistics control. The 
probability for node selection in ACA is displayed in 
equation (3). 
[ ( )] [ ( )]
,
() [[ ( )] [ ( )] ]
0,
ij ij
k
ij is is
tt
j unpassed
Pt tt
else







=




(3) 
In equation (3), 
ij

 is the information concentration. 
ij

 
is the path visibility. 

 is the pheromone trade-off 
factor. 

 is the heuristic factor for expected values. The 
ij

 is shown in equation (4). 
1
ij
ij
d
 =
           (4) 
In equation (4), 
ij
d
 is the distance between point i and 
point 
j
. After the ant colony completes the cycle, its 
pheromones will also be updated accordingly. The 
expression for pheromone update is shown in equation 
(5). 
( ) (1 ) ( ) ( )
ij ij ij
t n t t     + = − + 
   (5) 
In equation (5), 

 is the pheromone evaporation 
coefficient. t is the time point. However, traditional 
ACA tends to randomly select the next node. Although 
random selection helps explore larger task spaces, 
applying positive feedback also takes a long time, 
resulting in slower convergence speed in the initial stage. 
Neglecting the coordination and optimization of the entire 
supply chain can result in low operational efficiency of 
the entire logistics chain, increasing time cost. To this end, 
the study introduces logistic chaotic mapping, aiming to 
utilize the characteristics of logistic chaotic mapping to 
improve the accuracy of knowledge accumulation. Then, 
some randomness is added to the paths in the basic ACA 
during the optimization process. The principle of logistic 
chaotic mapping is shown in Figure 2. 
 
 
2.0 2.5 3.0 3.5 4.0 1.5 1.0 0.5 0
0
0.2
0.4
0.6
0.8
Control parameters
200 0
0
Iterations
Chaotic value
(a) Chaotic map
(b) Chaotic bifurcation diagram
Independent variable range
400 600 800 1000
0.2
0.4
0.6
0.8
 
Figure 2: Logistic chaos mapping 
 
Figure 2 (a) is a chaotic mapping diagram used to 
describe the evolution of the system's state over time 
under specific control parameters. It shows how the 
system state evolves over time given control parameters. 
In a chaotic state, the long-term behavior of the system is 
unpredictable. Even small initial condition changes can 
lead to completely different results. Figure 2 (b) is used to 
visualize the system behavior when the control 
parameters change. It shows the pattern of system 
behavior changing with control parameters and reveals 
the transition process from order to chaos. The mapping 
function for logistic chaotic mapping is shown in 

Automated Logistics Control Model Based on Improved Ant… Informatica 48 (2024) 13–26 17 
equation (6). 
1
4 (1 )
i i i
  
+
=−         (6) 
 
In equation (6), 

 is the chaotic invariant set. The 
characteristics of logical chaotic mapping can enable 
ACA to have a larger search range and randomness in the 
initial stage, which helps to escape from local optima and 
accelerate the search for global optima. Optimizing the 
initial pheromone distribution through logical chaotic 
mapping can make ants more diverse and randomized in 
path selection, which is beneficial for accelerating 
convergence speed and improving the search efficiency 
of the algorithm. Based on ACA, an automated logistics 
control system hardware framework is used to initialize 
and optimize the ant colony population, as shown in 
equation (7). 
( 1) ( )(1 ( )) t       + = −     (7) 
 
In equation (7), 

 is the number of iterations. 

 is the 
control parameter. In ACA, the initial distribution of 
pheromones has a significant impact on the performance. 
After some iterations, if the concentration of pheromones 
on the current optimal path is too high, traditional ACA 
may experience local optima. Therefore, the study aims 
to improve ACA by reducing excessive concentration of 
pheromones. The improved pheromone update 
mechanism is shown in equation (8). 
 
 
1
1
0
20
( 1) (1 ) ( )
,
1 ( ) ,
,
1 ( ) ,
ij ij ij
m
k
ij
k
ij
k
ij
tt
tT
a T t t T
tT
a t T t T
   






=


+ = −  + 





=

+ −   



 

=


+ −  
 

  (8) 
In equation (8), 
ij

 represents the pheromone 
concentration along the path from node i to 
j
, 
k
ij
  
represents the amount of pheromones released by ant k 
on path ( , ) ij. 

 represents the evaporation coefficient 
of pheromones. 
1
a and 
2
a are constants. Logistic 
chaotic mapping optimizes the initial pheromone 
distribution by randomly jumping the initial solution out 
of the local optimal solution domain, expanding the 
search range, and increasing the possibility of finding the 
global optimal solution. The improved algorithm 
flowchart is shown in Figure 3. 
 
Start
Initialization 
parameters
Initializing Ant 
Colony Individuals 
with Chaotic 
Mapping
Calculate the 
probability of ant 
transfer
Place ant nodes in 
the taboo list
Global pheromone 
update, limiting 
pheromone range
Place ant nodes in 
the taboo list
Output final 
optimization path
Finish
Is it the optimal 
solution
Has the maximum 
number of iterations 
been reached
Y
N
N
Y
 
Figure 3: Improved ant colony algorithm process 
 
Figure 3 showed the improved ACA process for this 
study. Firstly, the initial values of the chaotic map are set. 
Then the parameters of the ACA are set, including 
pheromone volatilization rate, pheromone intensity, ant 
population, etc. Next, logistic chaotic mapping is used to 
generate a set of new chaotic solutions. An improved 
ACA is used for searching. Then each ant updates its 
pheromones and selects the optimal solution based on the 
quality of the path during the search process. If the preset 
number of iterations is reached or a certain stopping 
criterion is met, the algorithm ends. Otherwise, the 
process returns to step 2 to continue iterating. In 
summary, this study introduces logistic chaotic mapping 
to improve the ACA, aiming to optimize the automated 
logistics system. 
3.2 Design of an automated logistics system 
model based on improved ant colony 
algorithm 
The automated logistics system model based on improved 
ACA mainly aims to achieve automation and 
optimization of the logistics process. Therefore, to 
improve the efficiency and accuracy of logistics systems, 
reduce cost, improve service quality, enhance scalability, 
and adapt to constantly changing market demands, an 
improved ACA is applied to design an automated 
logistics system model. In the model construction, the 
distribution path is first optimized. When the distribution 
source is 1 and the distribution task is one vehicle, the 
logistics optimization function is shown in equation (9). 

18   Informatica 48 (2024) 13–26                                                                    S. Ye
 
et al. 
 
1
1,
m
ik
k
y i W
=
=  

       (9) 
 
In equation (9), k is the refrigerated truck number. W 
is the set of target points. When the delivery source is 2, 
and the delivery task is one vehicle, the optimization 
function of the logistics path is shown in equation (10). 
,0
1,
m
ijk
ij
x k V
=
  
     (10) 
 
In equation (10), V represents the collection of 
refrigerated delivery vehicles. When the planned delivery 
route does not exceed the maximum travel distance of the 
delivery route, its expression is shown in equation (11). 
 
max
,0
m
ij ijn
ij
d x d n L
=
  
     (11) 
 
In equation (11), L is the delivery mileage. When the 
weight of the delivered goods does not exceed the 
maximum delivery capacity, the corresponding 
expression is shown in equation (12). 
0
,
m
i ik
i
q y Q k V
=
  

     (12) 
 
In equation (12), 
i
q represents the node demand. 
Q
 is 
the maximum load for logistics distribution. Automated 
logistics system refers to a system that achieves the 
automation and intelligence of logistics processes through 
advanced logistics technology and equipment, improving 
logistics efficiency and accuracy [20]. As an important 
equipment in automated logistics systems, stacker cranes 
are mainly used for the storage, retrieval, and sorting of 
goods. The picking operation is closely related to the 
automated logistics system. The operating trajectory of 
the stacker crane picking operation in this study is shown 
in Figure 4. 
y
x Column
Layer
Stacker
 
Figure 4: Stacker picking operation trajectory 
 
The picking operation of the designed stacker crane is 
divided into two strategies. One strategy is to randomly 
select storage locations for storage and retrieval 
operations. Based on this strategy, the running path of the 
stacker crane is randomly generated. A different path is 
selected for each run. The advantage of this strategy is 
high flexibility, which can adjust the running path at any 
time. Another strategy is to first select the first inventory 
task. Then the nearest actionable storage location is 
selected in sequence for operation until all task storage 
locations are traversed. According to this strategy, the 
running path of the stacker crane is generated 
sequentially based on the order of tasks. Each run selects 
the next nearest available storage location. The advantage 
of this strategy is that it can ensure that each task is 
traversed. The operational model of the time spent on 
stacker crane operations at different cargo points is 
shown in equation (13). 
 
 
max / , /
ij j i x j i y
T X X W V Y Y H V = −  − 
(13) 
In equation (13), X and Y are the coordinate points 
of goods transportation, respectively. The total time spent 
on stacker crane operations is shown in equation (14). 
 
1
,,
0
1
n
i i n o
i
T T T
−
=
= + +

      (14) 
 
In equation (14), 
n
 is the item number. Next, the path 
of the stacker crane is optimized to select the optimal 
sorting path to minimize the total time spent. The 
corresponding operating model is shown in equation (15). 
 
1
min , 1 ,0
0
min( ) min( )
n
i i n
i
T T T T
−
−
=
= = +

  (15) 
 
The picking operation of the stacker crane optimizes path 

Automated Logistics Control Model Based on Improved Ant… Informatica 48 (2024) 13–26 19 
planning, task allocation, operation sequence, and other 
aspects. This can make the operation of the stacker crane 
more efficient and stable, thereby improving the 
performance of the entire automated logistics system. The 
final designed automation hardware composition system 
is shown in Figure 5. 
Automated logistics system
Automated warehouse Control system Hybrid assembly line
Stacker
Double row 
combined shelves
Stacker control 
cabinet
Upper industrial 
control computer
Assembly line control 
cabinet
Entry and exit dock
Belt line
Double speed chain 
line
Roller line
Corner machine
 
Figure 5: Automated logistics control system 
 
Figure 5 shows the hardware framework of the automated 
logistics control system designed in this study. It mainly 
includes three parts: automated three-dimensional 
warehouse, control system, and hybrid assembly line. The 
improved ACA that integrates logistic chaotic mapping 
plays an important role in path planning and task 
scheduling. The main testing function used in system 
testing is shown in equation (16). 
2 2 2 2 2
1 2 1 2 2
22
2
max ( , ) 20( ) (1 )
3(1 ) 0.3
f x x x x x
x
= − − −
− − +
(16) 
By solving equation (16), the performance of the 
improved automatic logistics system can be tested. An 
automated three-dimensional warehouse is the core part 
of the automated logistics control system, consisting of 
high-rise shelves, stackers, handling robots, and inbound 
and outbound conveying systems. The control system is 
the core of the automated logistics control system, 
consisting of a control center, sensors, actuators, etc. 
Shelves are used to store goods. The stacker crane is 
responsible for storing and retrieving goods on the 
shelves. The handling robot is responsible for moving 
goods from one location to another. The inbound and 
outbound conveying system is used to deliver goods from 
the warehouse or from outside to the warehouse. The 
control center is responsible for receiving signals from 
sensors. According to the preset algorithm and program, 
the signal is processed and instructions are sent to the 
actuator. The actuator executes the corresponding actions. 
Hybrid assembly line is an efficient logistics 
transportation method, composed of various conveying 
equipment, such as belt conveyor, roller conveyor, chain 
conveyor, etc. In summary, the study aims to design an 
automated logistics control system based on an improved 
ACA, providing more accurate and timely decision 
support for logistics control.  
 
4 The application effect evaluation of 
improved ant colony algorithm in 
automated logistics control system 
To further verify the superiority of the improved ACA in 
the automated control systems and the performance of the 
designed automated logistics control system, the designed 
method is compared with other algorithms. 
 
4.1 Performance analysis of improved ant 
colony algorithm in automated logistics 
control model 
Logistic chaotic mapping is used to improve the ACA, 
and then an automated logistics control system is 
constructed. Firstly, the performance verification is 
performed on the proposed LACO algorithm. The 
experiment is conducted in MATLAB/Simulink for 
system simulation. By setting corresponding parameters 
and constraints, the performance of the LACO algorithm 
is observed. The experimental model parameters are set 
in Table 2. 
 
 

20   Informatica 48 (2024) 13–26                                                                    S. Ye
 
et al. 
 
 
 
Table 2: Basic parameters of experimental model 
Parameter Parameter units Parameter value 
Unit fixed cost Yuan 260 
Unit transport costs Yuan 3 
Vehicle speed km/h 55 
Maximum vehicle mileage limit km 12w 
Rated heavy load kg 4000 
Pallet sizes mm 460*450 
Number of positions 2 rows 5 floors * 15 columns 
 
The study is based on the Windows 11 platform, and the 
open-source MySQL software is used as the database. 
The number of projects is set to 50, the importance factor 
of the heuristic function is set to 5, the pheromone 
volatilization factor is set to 0.6, and the maximum 
number of iterations is set to 500. The DT100 dataset is 
used for testing. The classification results before and after 
the combination of LACO in this dataset are statistically 
analyzed. The DT100 dataset is a commonly used 
classification experiment dataset, consisting of a dataset 
of 100 nodes. It is used to test the performance of 
algorithms in dealing with large-scale problems. The 
node distribution of this dataset is relatively uniform, 
with high connectivity and density. Next, the 
performance of the improved ACA in the automated 
logistics control system is analyzed. Basic ACA, PSO, 
and SAA before improvement are selected for 
comparison. The loss curves of these four algorithms are 
compared. 
0
0.000
100 200 300 400 500 600 700 800
0.005
0.010
0.015
0.020
0.025
0.030
EI
Iterations
PSO
SAA
Basic ACA
Research algorithm
 
Figure 6: Comparison of loss curves of algorithms 
 
Figure 6 showed the comparison results of the loss curves 
for four algorithms. Compared with the other three 
algorithms, the proposed LACO algorithm could find the 
optimal solution after 200 iterations. The PSO algorithm 
had significant losses when iterated around 300 times. 
The corresponding basic ACA and SAA algorithms had 
significant fluctuations in their loss functions, with loss 
values ranging from 0.003 to 0.026. From this, the LACO 
algorithm has better superiority in dealing with complex 
problems. Because logistic chaotic mapping is used to 
generate initial solutions, it can guide ants to find the 
optimal solution faster. In addition, the LACO algorithm 
also utilized the advantages of ACA, such as swarm 
intelligence and distributed search, thereby improving 
search efficiency. Next, the cost solutions of the four 
algorithms are compared. The results are shown in Figure 
7. 
 

Automated Logistics Control Model Based on Improved Ant… Informatica 48 (2024) 13–26 21 
60 80 20 0
2.5
3.0
(a) PSO
2.0
1.5
Numerical value
Iterations
40 60 80 20 0
2.5
3.0
(b) SAA
2.0
1.5
Numerical value
Iterations
40
60 80 20 0
2.5
3.0
(c) Basic ACA
2.0
1.5
Numerical value
Iterations
40 60 80 20 0
2.5
3.0
(d) Research algorithm
2.0
1.5
Numerical value
Iterations
40
Minimum 
cost
Average cost
Minimum 
cost
Average cost
Minimum 
cost
Average cost
Minimum 
cost
Average cost
 
Figure 7: Comparison of algorithm cost solutions 
 
Figure 7 showed a comparison of cost solutions for four 
algorithms in automated logistics systems. The minimum 
objective function value when solving optimization 
problems is compared. Figure 7 (d) shows the cost 
comparison. The results showed that the proposed 
algorithm was 0.25 units lower than the average cost. The 
cost solutions of PSO algorithm, SAA algorithm, and 
basic ACA were all higher than the average cost. The 
average minimum cost of the research algorithm was 
about 0.51, 0.49, and 0.46 less than that of PSO, SAA, 
and basic ACA. From this, the LACO algorithm can 
continuously search and optimize the solution space by 
combining the advantages of logistic chaotic mapping 
and ACA, and quickly find cost solutions. Next, a 
comparative analysis is conducted on the distance 
solutions of the four algorithms. The analysis results are 
shown in Figure 8. 
 
60 80 20 0
2.5
3.0
(a) PSO
2.0
1.5
Iterations
40 60 80 20 0
2.5
3.0
(b) SAA
2.0
1.5
Iterations
40
60 80 20 0
2.5
3.0
(c) Basic ACA
2.0
1.5
Iterations
40 60 80 20 0
2.5
3.0
(d) Research algorithm
2.0
1.5
Iterations
40
Average distance
Closest distance
Average distance
Closest distance
Average distance
Closest distance
Average distance
Closest distance
Distance
Distance
Distance
Distance
 
Figure 8: Comparison of algorithm distance solutions 

22   Informatica 48 (2024) 13–26                                                                    S. Ye
 
et al. 
 
Figure 8 shows a comparison of distance solutions for 
four algorithms. The distance solution calculates the 
distance between the current solution and the optimal 
solution, and updates the pheromone and transition 
probability based on this distance. Figure 8 (d) shows the 
distance solution results of the algorithm in this study. 
Compared with the distance solutions of the other three 
algorithms, the distance solution of the research 
algorithm was overall depressed. The decrease was 
significant after 20 iterations. The research distance cost 
was 0.64, 0.54, and 0.51 less than PSO, SAA, and basic 
ACA, respectively. In summary, the LACO algorithm 
combines the advantages of logistic chaos mapping and 
ACA. It has low loss in automated logistics systems, 
which can quickly obtain the optimal solution, with 
strong global search ability. 
 
4.2 The effectiveness analysis of an 
automated logistics control model based on 
improved ant colony algorithm 
To verify the superiority of the designed automated 
logistics control system, real-time distribution paths are 
obtained through map APIs and sensor data in the 
experiment. The automated logistics control models 
based on PSO algorithm, SAA, basic ACA, and research 
algorithm are respectively used for comparing logistics 
distribution paths. The results are shown in Figure 9. 
 
5 10 15 20 25 30
5
10
15
20
25
30
35
Y coordinate
X coordinate
35
(a) PSO
5 10 15 20 25 30
5
10
15
20
25
30
35
Y coordinate
X coordinate
35
5 10 15 20 25 30
5
10
15
20
25
30
35
Y coordinate
X coordinate
35 5 10 15 20 25 30
5
10
15
20
25
30
35
Y coordinate
X coordinate
35
(b) SAA
(c) Basic ACA (d) Research algorithm
 
Figure 9: Comparison of logistics delivery path optimization models 
 
Figure 9 shows the distribution path results of different 
models under the same 14 distribution points. Figure 9 (d) 
shows the distribution path optimization results of this 
research model. Compared with the other three models, 
the logistics distribution path of the designed model did 
not exhibit node conflicts or path duplication. The total 
delivery distance was 12 km, 15 km, and 34 km less than 
the PSO algorithm, SAA, and basic ACA, respectively. 
From this, improving the ACA can optimize logistics 
distribution paths, reduce delivery time cost, and improve 
delivery efficiency. By simulating and optimizing the 
delivery process, the model can achieve faster and more 
accurate delivery services, meeting the timely needs of 
customers. Then, a comparison is made between the four 
models in terms of energy consumption. The comparison 
results are shown in Figure 10. 

Automated Logistics Control Model Based on Improved Ant… Informatica 48 (2024) 13–26 23 
5
0.4
10 15 20 25 30
0.5
0.6
0.7
0.8
0.9
1.0
Battery
Logistics node
PSO
SAA
Basic ACA
Research algorithm
 
Figure 10: Comparison of model energy consumption 
 
Figure 10 shows a comparison of model energy 
consumption. In the Figure, with the increase of time, the 
remaining electricity of the four logistics distribution 
models showed a decreasing trend. Compared with the 
other three models, the research model consumed less 
electricity. On average, each logistics node consumed 
1.01%, which was 0.21%, 0.67%, and 1.02% less than the 
PSO, SAA, and basic ACA, respectively. The model used 
in this study can reduce energy consumption and 
environmental pollution. Finally, the operational accuracy 
of the research model is compared. The results are shown 
in Figure 11. 
 
70 80 90 100
70
Predictive value
(a) Research algorithm
True value
80
90
100
70 80 90 100
70
Predictive value
(b) SAA
True value
80
90
100
R
2
=0.98 R
2
=0.71
70 80 90 100
70
Predictive value
(c) Basic ACA
True value
80
90
100
70 80 90 100
70
Predictive value
(d) PSO
True value
80
90
100
R
2
=0.68 R
2
=0.81
 
Figure 11: Comparison of model prediction accuracy 
 
Figure 11 compares the prediction accuracy of four 
different logistics models. Figure 11 (a) shows the 
prediction accuracy results of the automated logistics 
control model designed in this study. The results showed 
that the designed model had the highest prediction 
accuracy, with R2 of 0.98, which was 0.27, 0.30, and 
0.17 higher than the prediction accuracy of SAA, basic 
ACA, and PSO, respectively. In summary, the model can 
optimize logistics routes, and reduce energy consumption 
in logistics distribution, with high prediction accuracy. 
The fitting effect is better than other models. To further 
validate the reliability and scalability of the proposed 
model, the ApolloScape real-world dataset published by 
Baidu is used for testing. The accuracy and solution time 
of the proposed model are compared with SAA, basic 
ACA, and PSO algorithms. The results are shown in 

24   Informatica 48 (2024) 13–26                                                                    S. Ye
 
et al. 
Figure 12. 
 
(a) Accuracy
70
75
80
Accuracy/%
65
85
90
95
SAA ACA PSO
Research 
algorithms Mode
l
(b) Solution time
25
30
35
Solution time/s
20
40
45
SAA ACA PSO
Research 
algorithms
Mode
l
15
 
Figure 12: Comparison of accuracy and solution time of four models 
 
Figure 12 compares the accuracy and solution time of 
four different logistics models. Figure 12 (a) shows the 
accuracy comparison results. The proposed model had the 
highest solution accuracy of 92.58%. Figure 12 (b) shows 
the comparison of solution times. The solution time of the 
proposed model was 38.64s, which was higher than the 
other three models, but still within an acceptable range. 
5 Discussion 
To improve logistics efficiency, ACA was introduced 
into the logistics control model. Then the LACO 
algorithm was proposed to address the shortcomings of 
traditional ACA. Finally, an automated logistics system 
model was designed based on the improved ACA. The 
test results on the DT100 dataset showed that the LACO 
algorithm had strong optimization ability, which could 
find the optimal solution at 200 iterations, minimizing the 
objective function value to be more than 0.25 units lower 
than the average cost. Introducing logical chaotic 
mapping allows the algorithm to have a larger search 
range and randomness in the initial stage, making it more 
likely to find the global optimal solution. This 
improvement measure also enhanced the algorithm 
performance and to some extent solved the problems 
faced in references [8] and [10]. In addition, compared 
with the PSO, SAA, and basic ACA, the delivery distance 
of the proposed model was shorter, the average power 
consumption per logistics node was lower, at 1.01%, and 
the prediction accuracy was higher, with an R2 of 0.98. 
This indicates that the proposed model has good 
optimization ability, which can effectively improve 
delivery efficiency and reduce energy consumption, and 
has certain practical application value. Unlike references 
[11] and [13], the study conducted tests on real-world 
datasets to validate the performance of the proposed 
model. In the ApolloScape dataset, the solution time of 
the proposed model was 38.64s, which was higher than 
traditional ACA. Although this study improved the 
algorithm performance, it also increased the complexity 
of the algorithm, resulting in a decrease in its 
computational efficiency. This issue is similar to the 
problems in references [6], [9], [12], and [15]. Therefore, 
in future, it is necessary to further reduce the complexity 
of the algorithm as much as possible while ensuring 
performance. 
6 Conclusion 
With the continuous development of modern industrial 
automation technology, the logistics industry is 
undergoing tremendous changes. However, traditional 
automated logistics control systems often lack sufficient 
intelligence and adaptability. For this purpose, the study 
incorporated Logistic chaotic mapping into ACA to 
modify the logistics control system, verifying the 
superiority of the model through simulation experiments. 
Compared with the other three algorithms, the LACO 
algorithm could find the optimal solution at 200 iterations. 
The average minimum cost of research algorithms was 
about 0.51, 0.49, and 0.46 less than that of PSO, SAA, 
and basic ACA. The distance cost of research algorithm 
was 0.64, 0.54, and 0.51 less than that of PSO, SAA, and 
basic ACA, respectively. The total delivery distance of 
the research model was 12 km, 15 km, and 34 km less 
than that of the PSO, SAA, and basic ACA, respectively. 
The research model consumed less electricity, which was 
0.21%, 0.67%, and 1.02% less than the PSO, SAA, and 
basic ACA respectively. The highest prediction accuracy 
of the designed model was highest, with R2 of 0.98, 
which was 0.27, 0.30, and 0.17 higher than the prediction 
accuracy of SAA, basic ACA, and PSO, respectively. In 
summary, the designed logistics model can effectively 
improve the overall performance and efficiency of the 
logistics system. There are shortcomings in this study. 
The real-time traffic and speed of logistics distribution 
have not been considered. Further consideration can be 
given to the impact of real-time and unexpected situations 
in logistics distribution on the logistics system. 
 
 

Automated Logistics Control Model Based on Improved Ant… Informatica 48 (2024) 13–26 25 
References 
[1] Q. Han, Y. Sun, Q. Wu, and Z. Bai, “Research on 
optimization model of logistics transportation truck 
path considering environmental impact: 
experimental data from xiqing district, Tianjin,” 
Journal of Advanced Transportation, vol. 2, no. 5, 
pp. 786-796, 2021. 
https://doi.org/10.1155/2021/6665168 
[2] A. Delgoshaei, H. Norozi, A. Mirzazadeh, M. 
Farhadi, G. H. Pakdel, and A. K. Aram, “A new 
model for logistics and transportation of fashion 
goods in the presence of stochastic market demands 
considering restricted retailers capacity,” 
RAIRO-Operations Research, vol. 55, no. 9, pp. 
523-547, 2021. https://doi.org/10.1051/ro/2019061 
[3] X. Wang, N. Guan, and J. Yang, “Image encryption 
algorithm with random scrambling based on 
one-dimensional logistic self-embedding chaotic 
map,” Chaos Solitons & Fractals, vol. 150, no. 3, pp. 
111-117, 2021. 
https://doi.org/10.1016/j.chaos.2021.111117 
[4] W. Hwang, D. Kang, and D. Kim. “Brain 
lateralisation feature extraction and ant colony 
optimisation-bidirectional LSTM network model for 
emotion recognition,” IET Signal Process, vol. 16, 
pp. 45-61, 2021. https://doi.org/10.1049/sil2.12076 
[5] T. R. Brodetzki, G. Brodetzki, O. Feinerman, and A. 
Hefetz, “Worker demography and behavior in a 
supercolonial ant colony: The case of the desert ant 
Cataglyphis niger,” Ethology, vol. 8, no. 3, pp. 
876-889, 2020.  https://doi.org/10.1111/eth.12960 
[6] A. F. Güven, N. Yrükeren, and M. M. Samy, 
“Design optimization of a stand-alone green energy 
system of university campus based on jaya-harmony 
search and ant colony optimization algorithms 
approaches,” Energy, vol. 253, no. 7, pp. 675-687, 
2022. https://doi.org/10.1016/j.energy.2022.124089 
[7] B. Moradi, A. Kargar, and S. Abazari, “Transient 
stability constrained optimal power flow solution 
using ant colony optimization for continuous 
domains (ACO_R),” IET Generation, Transmission 
& Distribution, vol. 7, no. 9, pp. 786-798, 2022. 
https://doi.org/10.1049/gtd2.12560 
[8] C. Li, K. Jiang, and Y. Luo, “Dynamic placement of 
multiple controllers based on SDN and allocation of 
computational resources based on heuristic ant 
colony algorithm,” Knowledge-Based Systems, vol. 
241, no. 7, pp. 108-119, 2022. 
https://doi.org/10.1016/j.knosys.2022.108330 
[9] W. He, S. Meng, J. Wang, L. Wang, R. Pan, and W. 
Gao, “Weaving scheduling based on an improved 
ant colony algorithm,” Textile Research Journal, vol. 
91, no. 6, pp. 543-554, 2021. 
https://doi.org/10.1177/0040517520948896 
[10] A. S. D. Pontes, A. S. D. Araújo, and W. Marinho, 
“Ant colony optimization for variable selection in 
discriminant linear analysis,” Journal of 
Chemometrics, vol. 77, no. 4, pp.897-903, 2020. 
https://doi.org/10.1002/cem.3292 
[11] Y. Zhu, and J. Suo, “The situation of offshore oil 
logistics industry and transportation scheme design 
of new concept of green logistics,” Journal of 
Coastal Research, vol. 112, no. 12, pp. 112-123, 
2020. https://doi.org/10.2112/JCR-SI112-118.1 
[12] H. Jin, Q. He, M. He, S. Lu, F. Hu, and D. Hao, 
“Optimization for medical logistics robot based on 
model of traveling salesman problems and vehicle 
routing problems,” International Journal of 
Advanced Robotic Systems, vol. 18, no. 3, pp. 
103-121, 2021. 
https://doi.org/10.1177/17298814211022539 
[13] X. Zhang, “B2C E-commerce logistics network 
optimization model,” Journal of Global Information 
Management, vol. 6, no. 11, pp. 245-254, 2022. 
https://doi.org/10.4018/JGIM.20220701.oa7 
[14] C. M. Defalque, A. F. D. Silva, and F A S. Marins, 
“Goal programming model applied to waste paper 
logistics processes,” Applied Mathematical 
Modelling, vol. 7, no. 3, pp. 786-792, 2021. 
https://doi.org/10.1016/j.apm.2021.05.002 
[15] X. Guo, W. Zhang, and B. Liu, “Low-carbon 
routing for cold-chain logistics considering the 
time-dependent effects of traffic congestion,” 
Transportation Research Part D: Transport and 
Environment, vol. 6, no. 5, pp. 332-345, 2022. 
https://doi.org/10.1016/j.trd.2022.103502 
[16] S. Kumar, B. Sarkar, and A. Kumar, “Fuzzy reverse 
logistics inventory model of smart items with two 
warehouses of a retailer considering carbon 
emissions,” RAIRO - Operations Research, vol. 55, 
no. 4, pp. 2285-2307, 2021. 
https://doi.org/10.1051/ro/2021056 
[17] S. Sharma, S. R. Singh, and M. Kumar, “A reverse 
logistics inventory model with multiple production 
and remanufacturing batches under fuzzy 
environment,” RAIRO - Operations Research, vol. 
55, no. 2, pp. 571-588, 2021. 
https://doi.org/10.1051/ro/2021021 
[18] Z. Wang, S. Gao, Y. Zhang, and L. Guo, 
“Symmetric uncertainty-incorporated probabilistic 
sequence-based ant colony optimization for feature 
selection in classification,” Knowledge-based 
Systems, vol. 7, no. 4, pp. 78-86, 2022. 
https://doi.org/10.1016/j.knosys.2022.109874 
[19] D. Zhao, L. Liu, F. Yu, A. A. Heidari, and H. Chen, 
“Chaotic random spare ant colony optimization for 
multi-threshold image segmentation of 2D Kapur 
entropy,” Knowledge-Based Systems, vol. 7, no. 5, 
pp. 234-245, 2020. 
https://doi.org/10.1016/j.knosys.2020.106510 
[20] P. P. Groumpos, “A critical historic overview of 
artificial intelligence: issues, challenges, 
opportunities, and threats,” Artificial Intelligence 
and Applications. Vol. 1, no. 4, pp. 197-213, 2023. 

26   Informatica 48 (2024) 13–26                                                                    S. Ye
 
et al. 
https://doi.org/10.47852/bonviewAIA3202689