Logistics sorting and carrying system integrated optimization method

Abstract

The invention relates to an integrated optimization method for a logistics sorting and carrying system, and belongs to the technical field of automatic logistics sorting systems. According to the actual operation condition of the logistics sorting and carrying system, firstly, an integrated optimization problem model of the logistics sorting and carrying system is established; secondly, constructing a simulation optimization framework for solving the problem model; then, designing a multi-target whale optimization algorithm as an optimization algorithm module of the simulation optimization framework; and finally, designing an improved A * algorithm as a path planning module of the simulation optimization framework. A decision scheme of sorting task distribution and AGV path planning is obtained through simulation optimization, and decision basis is provided for management personnel. The method is established based on the actual working condition of the general logistics sorting and carrying system, the proposed method is close to reality, the sorting time can be effectively shortened, and the AGV operation efficiency is improved.

Description

Technical Field

[0001] The invention is used for a logistics sorting and handling system integration optimization method and belongs to the technical field of automated logistics sorting systems. Background Art

[0002] With the rapid development of sorting technology in the logistics industry, logistics sorting and handling systems, as an important sorting method, have been widely used in related industries. Automated guided vehicles (AGVs) based on autonomous positioning and navigation technologies have gradually become the best choice for logistics sorting and handling equipment due to their automation, high flexibility, high efficiency, high reliability, and parallel operation. The logistics sorting and handling system with AGV as the intermediate handling equipment has a higher level of flexibility and automation than the traditional system. However, the coupling relationship between sorting scheduling and AGV path planning in the logistics sorting and handling system also brings challenges to managers' decision-making.

[0003] At present, most studies on the integrated optimization of logistics sorting and handling systems optimize sorting scheduling and AGV path planning as independent parts. However, due to the coupling relationship between the two, independent optimization has a limited effect on improving the overall efficiency of the system. In view of the shortcomings of current research on logistics sorting and handling systems, the present invention proposes an integrated optimization method for logistics sorting and handling systems, which comprehensively considers the impact of sorting scheduling and AGV path planning on the efficiency of logistics sorting and handling systems. Summary of the invention

[0004] The purpose of the present invention is to propose a method for optimizing the integration of a logistics sorting and handling system in view of the shortcomings of the existing integrated optimization of logistics sorting and handling systems.

[0005] The technical solution adopted by the present invention to solve the technical problem is: first, establish a logistics sorting and handling system integrated optimization problem model; second, construct a simulation optimization framework for solving the problem model; then, design a multi-objective whale optimization algorithm as an optimization algorithm module of the simulation optimization framework; finally, design an improved A* algorithm as a path planning module of the simulation optimization framework;

[0006] The establishment of the integrated optimization problem model of the logistics sorting and handling system is described in detail as follows:

[0007] Based on the principle of shortest total AGV driving distance, the first optimization goal is established as follows:

[0008]

[0009] Where F1 represents the sum of the AGV travel distances; i∈{1,2,…,I} represents the sorting task index; I represents the total number of sorting tasks; j∈{1,2,…,J} represents the AGV index; J represents the total number of AGVs; U ijL represents the empty travel distance of AGV j handling and sorting task i; ij It represents the travel distance of AGVj in carrying and sorting task i.

[0010] Based on the principle of shortest completion time for the largest sorting task, the second optimization goal is established as follows:

[0011]

[0012] Where, F2 represents the maximum sorting task completion time; T i It indicates the time when sorting task i is completed and transported to the unloading site.

[0013] At the same time, the problem model also needs to meet the following constraints:

[0014] U ij =v×(TU′ ij -TU ij -n ij ×TW); where TU ij represents the empty-load start time of AGVj handling and sorting task i; TU′ ij represents the empty-load end time of AGVj handling and sorting task i; TW represents the turning time of AGV; n ij represents the number of turns made by AGVj in handling and sorting task i; v represents the driving speed of AGV.

[0015] L ij =v×(TL' ij -TL ij -n ij ×TW); where TL ij represents the load start time of AGVj carrying sorting task i; TL' ij Indicates the load completion time of AGVj's handling and sorting task i.

[0016] T i =E ik +TL' ij -TL ij ; In the formula, k∈{1,2,…,K}, represents the sorting station index; K represents the total number of sorting stations; E ik It represents the sorting completion time of sorting task i at sorting station k.

[0017] In the formula, X ik is a decision variable, indicating whether sorting task i is processed at sorting station k; this formula ensures that any sorting task is processed at only one sorting station.

[0018] E ik ≤S pk+M(1-Y ipk )i,p∈I; where Y ipk is a decision variable, indicating that sorting task i is processed at sorting station k before sorting task p; S pk It represents the start sorting time of sorting task p at sorting station k; M is an infinite integer; this formula ensures that the start sorting time of the sorting task at the same sorting station shall not be earlier than the sorting completion time of the previous sorting task.

[0019] In the formula, Z ij is a decision variable, indicating whether sorting task i is carried out by AGVj; this formula ensures that any sorting task is carried out by only one AGV.

[0020] TU ij ≤TL pj +M(1-W ijp )i,p∈I; where W ijp is a decision variable, indicating that sorting task i is carried out by AGVj before sorting task p; TL' pj It represents the load end time of AGVj's transport sorting task p; this formula ensures that the empty-load start time of the AGV's transport sorting task shall not be earlier than the load end time of the previous sorting task it transports.

[0021] The simulation optimization framework for solving the problem model is specifically described as follows:

[0022] The framework consists of four modules: control model, simulation model, optimization algorithm and path planning. When the system starts running, the control model module inputs parameters, and passes the model parameters and algorithm parameters to the simulation model module and the optimization algorithm module for model parameter setting and algorithm parameter setting respectively. The simulation model module builds the simulation model according to the model parameters. The optimization algorithm module initializes the parameters according to the algorithm parameters and generates the initial population. The initial population is fed back to the control model module, the population individuals are decoded and the sorting task allocation plan is obtained. The allocation plan is fed back to the simulation model module to run the simulation model. When running the simulation model, the path planning module is called to obtain the starting and ending positions of the sorting task according to the sorting task allocation plan, and the AGV path planning plan is obtained by improving the A* algorithm as the AGV planning path for the handling and sorting task. At the end of the simulation, the simulation results are fed back to the optimization algorithm module for non-dominated sorting and the algorithm search operation is completed. In the optimization algorithm module, it is determined whether the termination condition is met: if it is met, the result is output to obtain the optimal sorting task allocation and AGV path planning decision plan, providing decision-making basis for managers; otherwise, the new population individuals are fed back to the control model module for decoding to continue the algorithm iteration optimization.

[0023] The multi-objective whale optimization algorithm is designed as the optimization algorithm module of the simulation optimization framework. The multi-objective whale optimization algorithm is specifically described as follows:

[0024] Step 1, initialization parameters;

[0025] Step 2, generate the initial population, using a hybrid initialization method that combines a random strategy and a heuristic strategy based on balanced sorting time;

[0026] Step 3: decoding, the individuals in the population decode in turn to obtain the sorting task allocation plan;

[0027] Step 4: non-dominated sorting. Run the simulation model with the sorting task allocation plan as a precondition. After the simulation, get the individual fitness function value, perform Pareto non-dominated sorting on the individuals, write the most cutting-edge solution set into the elite pool, perform Pareto non-dominated sorting on the individuals in the elite pool, and retain the best individuals in the elite pool.

[0028] Step 5, generate a random number p in the interval [0,1]. If p is less than or equal to 0.5, go to step 6; otherwise, go to step 9;

[0029] Step 6, calculate coefficient A, if A is less than or equal to 1, go to step 7; otherwise, go to step 8;

[0030] Step 7, elite individual crossover, the current individual is crossovered with the randomly selected elite individual;

[0031] Step 8, another individual crossover, the current individual is crossovered with another randomly selected individual;

[0032] Step 9, the current individual mutates, and the current individual performs mutation operation alone;

[0033] Step 10, determine whether the termination condition is met, if so, go to step 11; otherwise, go to step 3 to continue algorithm iterative optimization;

[0034] Step 11, output the optimal solution set and the algorithm terminates.

[0035] The improved A* algorithm is designed as the path planning module of the simulation optimization framework. The improved A* algorithm is specifically described as follows:

[0036] Step 1: Obtain location information and obtain the starting and ending locations of the current sorting task according to the sorting task allocation plan;

[0037] Step 2, initialize the list, define the Open list for storing the nodes to be expanded and the Close list for storing the expanded nodes, both lists are initialized to empty lists, add the starting node to the Open list and set its heuristic function value to 0;

[0038] Step 3: Find a reachable node, select the node with the smallest heuristic function value in the Open list, and find the reachable node of the node;

[0039] Step 4, expand the node, remove the selected node from the Open list, and add it to the Close list;

[0040] Step 5, calculate the function value, and calculate the function value of all reachable nodes according to the heuristic function;

[0041] Step 6, determine whether the reachable node is in the Open list or the Close list; if the reachable node is in the Close list, remove the reachable node; otherwise, go to step 7;

[0042] Step 7, determine whether the reachable node is in the Open list; if the reachable node is in the Open list, update the Open list according to the principle of minimum function value; otherwise, add the reachable node to the Open list;

[0043] Step 8, determine whether the end point location node is added to the Open list; if it is added to the list, it means that the path has been found, and the path planning solution is output to end the search process; otherwise, return to step 3 to continue searching.

[0044] The beneficial effects of the present invention are:

[0045] 1. Different from the traditional decision-making scheme, the present invention integrates and optimizes the logistics sorting and handling system to fit the actual production;

[0046] 2. The integrated optimization problem model of the logistics sorting and handling system established by the present invention optimizes multiple objectives simultaneously;

[0047] 3. The multi-objective whale optimization algorithm proposed in the present invention can adaptively adjust and improve the model and obtain a better optimal solution set;

[0048] 4. The model constructed by the present invention is targeted at the actual operating conditions of the general logistics sorting and handling system, has good practicality, and can be widely used in relevant systems of various industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described below with reference to the accompanying drawings, in which:

[0050] Figure 1 It is the overall flow chart of the present invention;

[0051] Figure 2 It is a flow chart of the simulation optimization framework of the present invention;

[0052] Figure 3 This is a flow chart of the multi-objective whale optimization algorithm of the present invention;

[0053] Figure 4 Improved A* algorithm flow chart for the present invention;

[0054] Figure 5 It is a topological map of the logistics sorting and handling system of the present invention. DETAILED DESCRIPTION

[0055] Example 1: Figure 1-5 As shown in the figure, the integrated optimization method of the logistics sorting and handling system is as follows: first, a logistics sorting and handling system integrated optimization problem model is established; second, a simulation optimization framework for solving the problem model is constructed; then, a multi-objective whale optimization algorithm is designed as the optimization algorithm module of the simulation optimization framework; finally, an improved A* algorithm is designed as the path planning module of the simulation optimization framework;

[0056] The establishment of the integrated optimization problem model of the logistics sorting and handling system is described in detail as follows:

[0057] Based on the principle of shortest total AGV driving distance, the first optimization goal is established as follows:

[0058]

[0059] Where F1 represents the sum of the AGV travel distances; i∈{1,2,…,I} represents the sorting task index; I represents the total number of sorting tasks; j∈{1,2,…,J} represents the AGV index; J represents the total number of AGVs; U ij L represents the empty travel distance of AGV j handling and sorting task i; ij It represents the travel distance of AGVj in carrying and sorting task i.

[0060] Based on the principle of shortest completion time for the largest sorting task, the second optimization goal is established as follows:

[0061]

[0062] Where, F2 represents the maximum sorting task completion time; T i It indicates the time when sorting task i is completed and transported to the unloading site.

[0063] At the same time, the problem model also needs to meet the following constraints:

[0064] U ij =v×(TU′ ij -TU ij -n ij ×TW); where TU ij represents the empty-load start time of AGVj handling and sorting task i; TU′ij represents the empty-load end time of AGVj handling and sorting task i; TW represents the turning time of AGV; n ij represents the number of turns made by AGVj in handling and sorting task i; v represents the driving speed of AGV.

[0065] L ij =v×(TL' ij -TL ij -n ij ×TW); where TL ij represents the load start time of AGVj carrying sorting task i; TL' ij Indicates the load completion time of AGVj's handling and sorting task i.

[0066] T i =E ik +TL' ij -TL ij ; In the formula, k∈{1,2,…,K}, represents the sorting station index; K represents the total number of sorting stations; E ik It represents the sorting completion time of sorting task i at sorting station k.

[0067] In the formula, X ik is a decision variable, indicating whether sorting task i is processed at sorting station k; this formula ensures that any sorting task is processed at only one sorting station.

[0068] E ik ≤S pk +M(1-Y ipk )i,p∈I; where Y ipk is a decision variable, indicating that sorting task i is processed at sorting station k before sorting task p; S pk It represents the start sorting time of sorting task p at sorting station k; M is an infinite integer; this formula ensures that the start sorting time of the sorting task at the same sorting station shall not be earlier than the sorting completion time of the previous sorting task.

[0069] In the formula, Z ij is a decision variable, indicating whether sorting task i is carried out by AGVj; this formula ensures that any sorting task is carried out by only one AGV.

[0070] TU ij ≤TL pj +M(1-W ijp )i,p∈I; where W ijp is a decision variable, indicating that sorting task i is carried out by AGVj before sorting task p; TL' pjIt represents the load end time of AGVj's transport sorting task p; this formula ensures that the empty-load start time of the AGV's transport sorting task shall not be earlier than the load end time of the previous sorting task it transports.

[0071] The simulation optimization framework for solving the problem model is specifically described as follows:

[0072] The framework consists of four modules: control model, simulation model, optimization algorithm and path planning. When the system starts running, the control model module inputs parameters, and passes the model parameters and algorithm parameters to the simulation model module and the optimization algorithm module for model parameter setting and algorithm parameter setting respectively. The simulation model module builds the simulation model according to the model parameters. The optimization algorithm module initializes the parameters according to the algorithm parameters and generates the initial population. The initial population is fed back to the control model module, the population individuals are decoded and the sorting task allocation plan is obtained. The allocation plan is fed back to the simulation model module to run the simulation model. When running the simulation model, the path planning module is called to obtain the starting and ending positions of the sorting task according to the sorting task allocation plan, and the AGV path planning plan is obtained by improving the A* algorithm as the AGV planning path for the handling and sorting task. At the end of the simulation, the simulation results are fed back to the optimization algorithm module for non-dominated sorting and the algorithm search operation is completed. In the optimization algorithm module, it is determined whether the termination condition is met: if it is met, the result is output to obtain the optimal sorting task allocation and AGV path planning decision plan, providing decision-making basis for managers; otherwise, the new population individuals are fed back to the control model module for decoding to continue the algorithm iteration optimization.

[0073] The multi-objective whale optimization algorithm is designed as the optimization algorithm module of the simulation optimization framework. The multi-objective whale optimization algorithm is specifically described as follows:

[0074] According to the integrated optimization problem model of the logistics sorting and handling system established in this goal, the multi-objective whale optimization algorithm proposed in this invention designs the encoding and decoding method, the initial population generation method, the algorithm search operation, etc., and optimizes through Pareto non-dominated sorting to obtain the optimal solution set. The multi-objective whale optimization algorithm includes the following steps:

[0075] Step 1, initialization parameters;

[0076] The multi-objective whale optimization algorithm population individuals use integer coding based on sorting tasks, and the total length of the coding is equal to the number of sorting tasks. During decoding, the sorting tasks are assigned to the corresponding sorting stations according to the coding, the sorting task allocation plan is obtained, and the AGV handling sorting tasks are selected based on the principle of minimum empty travel.

[0077] Step 2, generate the initial population, using a hybrid initialization method that combines a random strategy and a heuristic strategy based on balanced sorting time;

[0078] The quality of the initial solution has a significant impact on the performance of the algorithm. Relying solely on the heuristic strategy based on balanced sorting time can ensure the quality of the initial solution, but it will reduce the diversity of the population. In order to balance quality and diversity, a mixed initialization strategy is adopted, that is, 50% of the individuals are generated using the heuristic initialization strategy, and the other 50% of the individuals are generated using the random initialization strategy.

[0079] Step 3: decoding, the individuals in the population decode in turn to obtain the sorting task allocation plan;

[0080] Step 4: non-dominated sorting. Run the simulation model with the sorting task allocation plan as a precondition. After the simulation, get the individual fitness function value, perform Pareto non-dominated sorting on the individuals, write the most cutting-edge solution set into the elite pool, perform Pareto non-dominated sorting on the individuals in the elite pool, and retain the best individuals in the elite pool.

[0081] Step 5, generate a random number p in the interval [0,1]. If p is less than or equal to 0.5, go to step 6; otherwise, go to step 9;

[0082] Step 6, calculate coefficient A. If A is less than or equal to 1, go to step 7; otherwise, go to step 8. Coefficient A is expressed as:

[0083] A=a×(2×r1-1)

[0084] Where r1 is a random number generated in the interval [0,1]; a is the convergence factor that decreases linearly from 2 to 0 with the number of iterations.

[0085] a=2-2×t / t max

[0086] Where t is the current iteration number; t max is the maximum number of iterations.

[0087] Step 7, elite individual crossover, the current individual is crossovered with the randomly selected elite individual;

[0088] Step 8, another individual crossover, the current individual is crossovered with another randomly selected individual;

[0089] Step 9, the current individual mutates, and the current individual performs mutation operation alone;

[0090] Step 10, determine whether the termination condition is met, if so, go to step 11; otherwise, go to step 3 to continue algorithm iterative optimization;

[0091] Step 11, output the optimal solution set and the algorithm terminates.

[0092] The improved A* algorithm is designed as the path planning module of the simulation optimization framework. The improved A* algorithm is specifically described as follows:

[0093] Step 1: Obtain location information and obtain the starting and ending locations of the current sorting task according to the sorting task allocation plan;

[0094] Step 2, initialize the list, define the Open list for storing the nodes to be expanded and the Close list for storing the expanded nodes, both lists are initialized to empty lists, add the starting node to the Open list and set its heuristic function value to 0;

[0095] Step 3: Find a reachable node, select the node with the smallest heuristic function value in the Open list, and find the reachable node of the node;

[0096] Step 4, expand the node, remove the selected node from the Open list, and add it to the Close list;

[0097] Step 5: Calculate the function value. Calculate the function value of all reachable nodes according to the heuristic function. The heuristic function is expressed as:

[0098] f(n)=g(n)+h(n)+j(n)

[0099] Where f(n) is the approximate cost from the starting node to the end node via the reachable node n; g(n) is the actual distance from the starting node to the reachable node n; h(n) is the estimated distance from the reachable node n to the end node; j(n) is the turning penalty cost of extending the reachable node n.

[0100] j(n)=α×[(Y c -Y s )×(X c -X s )]

[0101] Where, X c and Y c is the x-coordinate and y-coordinate of the current node; s and Y s are the x- and y-coordinates of the reachable nodes; α is the turning penalty factor.

[0102] Step 6, determine whether the reachable node is in the Open list or the Close list; if the reachable node is in the Close list, remove the reachable node; otherwise, go to step 7;

[0103] Step 7, determine whether the reachable node is in the Open list; if the reachable node is in the Open list, update the Open list according to the principle of minimum function value; otherwise, add the reachable node to the Open list;

[0104] Step 8, determine whether the end point location node is added to the Open list; if it is added to the list, it means that the path has been found, and the path planning solution is output to end the search process; otherwise, return to step 3 to continue searching.

[0105] Problem setting: Figure 5 This is a topological map of a logistics sorting and handling system with AGV as the intermediate handling equipment. The system has a total of 16 sorting stations, 5 unloading stations and 12 charging stations. There are 12 AGVs in total, the speed of the AGV is 1m / s, and the turning time is 10s. The example is randomly generated. At the initial moment, all AGVs are parked at the charging station; after the sorting task is generated, the sorting is performed at the corresponding sorting station according to the sorting task allocation scheme obtained by coding; after the sorting is completed, the AGV is selected to transport the sorting task to the unloading station according to the principle of minimum empty travel; the AGV that has completed the handling returns to the charging station to dock.

[0106] Algorithm parameter setting: The algorithm parameters were determined through orthogonal experiments, where population size = 50 and maximum number of iterations = 100;

[0107] The following table shows the Pareto solution set obtained by this method.

[0108]

[0109] The specific embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A logistics sorting and handling system integration optimization method, characterized by: Firstly, the integrated optimization problem model of the logistics sorting and handling system is established; secondly, a simulation optimization framework for solving the problem model is constructed; then, a multi-objective whale optimization algorithm is designed as the optimization algorithm module of the simulation optimization framework; finally, an improved A* algorithm is designed as the path planning module of the simulation optimization framework; The establishment of the integrated optimization problem model of the logistics sorting and handling system is described in detail as follows: Based on the principle of shortest total AGV driving distance, the first optimization goal is established as follows: Where F1 represents the sum of the AGV travel distances; i∈{1,2,…,I} represents the sorting task index; I represents the total number of sorting tasks; j∈{1,2,…,J} represents the AGV index; J represents the total number of AGVs; U ij L represents the empty travel distance of AGV j handling and sorting task i; ij It represents the travel distance of AGVj carrying and sorting task i; Based on the principle of shortest completion time for the largest sorting task, the second optimization goal is established as follows: Where, F2 represents the maximum sorting task completion time; T i Indicates the time when sorting task i is completed and transported to the unloading site; At the same time, the problem model also needs to meet the following constraints: U ij =v×(TU i ' j -TU ij -n ij ×TW); where TU ij represents the empty-load start time of AGVj handling and sorting task i; TU i ' j represents the empty-load end time of AGVj handling and sorting task i; TW represents the turning time of AGV; n ij represents the number of turns of AGVj in handling and sorting task i; v represents the driving speed of AGV; L ij =v×(TL' ij -TL ij -n ij ×TW); where TL ij represents the load start time of AGVj carrying sorting task i; TL' ij Indicates the load completion time of AGVj handling and sorting task i; T i =E ik +TL' ij -TL ij ; In the formula, k∈{1,2,…,K}, represents the sorting station index; K represents the total number of sorting stations; E ik represents the sorting completion time of sorting task i at sorting station k; Where, X ik is a decision variable, indicating whether sorting task i is processed at sorting station k; this formula ensures that any sorting task is processed at only one sorting station; E ik ≤S pk +M(1-Y ipk )i,p∈I; where Y ipk is a decision variable, indicating that sorting task i is processed at sorting station k before sorting task p; S pk represents the start sorting time of sorting task p at sorting station k; M is an infinite integer; this formula ensures that the start sorting time of the sorting task at the same sorting station shall not be earlier than the sorting completion time of the previous sorting task; In the formula, Z ij is a decision variable, indicating whether sorting task i is carried by AGVj; this formula ensures that any sorting task is carried by only one AGV; TU ij ≤TL pj +M(1-W ijp )i,p∈I; where W ijp is a decision variable, indicating that sorting task i is carried out by AGVj before sorting task p; TL' pj It represents the load end time of AGVj's transport sorting task p; this formula ensures that the empty-load start time of the AGV's transport sorting task shall not be earlier than the load end time of the previous sorting task it transports; The simulation optimization framework for solving the problem model is specifically described as follows: The framework consists of four modules: control model, simulation model, optimization algorithm and path planning. When the system starts running, the control model module inputs parameters, and passes the model parameters and algorithm parameters to the simulation model module and the optimization algorithm module for model parameter setting and algorithm parameter setting respectively. The simulation model module builds the simulation model according to the model parameters. The optimization algorithm module initializes the parameters according to the algorithm parameters and generates the initial population. The initial population is fed back to the control model module, the individuals in the population are decoded and the sorting task allocation plan is obtained. Feed the allocation plan back to the simulation model module to run the simulation model; call the path planning module when running the simulation model, obtain the starting and ending positions of the sorting task according to the sorting task allocation plan, and obtain the AGV path planning plan for the AGV planning path of the handling and sorting task by improving the A* algorithm; after the simulation, feed the simulation results back to the optimization algorithm module for non-dominated sorting and complete the algorithm search operation; in the optimization algorithm module, determine whether the termination condition is met: if it is met, output the result, and obtain the optimal sorting task allocation and AGV path planning decision plan to provide decision-making basis for managers; otherwise, feed the new population individuals back to the control model module for decoding to continue algorithm iterative optimization; The multi-objective whale optimization algorithm is designed as the optimization algorithm module of the simulation optimization framework. The multi-objective whale optimization algorithm is specifically described as follows: Step 1, initialization parameters; Step 2, generate the initial population, using a hybrid initialization method that combines a random strategy and a heuristic strategy based on balanced sorting time; Step 3: decoding, the individuals in the population decode in turn to obtain the sorting task allocation plan; Step 4: non-dominated sorting. Run the simulation model with the sorting task allocation plan as a precondition. After the simulation, get the individual fitness function value, perform Pareto non-dominated sorting on the individuals, write the most cutting-edge solution set into the elite pool, perform Pareto non-dominated sorting on the individuals in the elite pool, and retain the best individuals in the elite pool. Step 5, generate a random number p in the interval [0,1]. If p is less than or equal to 0.5, go to step 6; Otherwise, go to step 9; Step 6, calculate coefficient A, if A is less than or equal to 1, go to step 7; Otherwise, go to step 8; Step 7, elite individual crossover, the current individual is crossovered with the randomly selected elite individual; Step 8, another individual crossover, the current individual is crossovered with another randomly selected individual; Step 9, the current individual mutates, and the current individual performs mutation operation alone; Step 10, determine whether the termination condition is met, if so, go to step 11; Otherwise, go to step 3 to continue algorithm iterative optimization; Step 11, output the optimal solution set and the algorithm terminates; The improved A* algorithm is designed as the path planning module of the simulation optimization framework. The improved A* algorithm is specifically described as follows: Step 1: Obtain location information and obtain the starting and ending locations of the current sorting task according to the sorting task allocation plan; Step 2, initialize the list, define the Open list for storing the nodes to be expanded and the Close list for storing the expanded nodes, both lists are initialized to empty lists, add the starting node to the Open list and set its heuristic function value to 0; Step 3: Find a reachable node, select the node with the smallest heuristic function value in the Open list, and find the reachable node of the node; Step 4, expand the node, remove the selected node from the Open list, and add it to the Close list; Step 5, calculate the function value, and calculate the function value of all reachable nodes according to the heuristic function; Step 6, determine whether the reachable node is in the Open list or the Close list; If the reachable node is in the Close list, remove the reachable node; Otherwise, go to step 7; Step 7, determine whether the reachable node is in the Open list; If the reachable node is in the Open list, update the Open list based on the principle of minimum function value; Otherwise, add the reachable node to the Open list; Step 8, determine whether the end location node is added to the Open list; If it is added to the list, it means that the path has been found, and the path planning solution is output to end the search process; Otherwise, return to step 3 to continue searching.

Images