Task scheduling method of intelligent warehousing system

Abstract

The invention relates to the technical field of intelligent warehousing system control, in particular to a task scheduling method of an intelligent warehousing system. The method comprises the following steps: firstly, generating a warehousing system environment map and a general task pool according to real-time data of order information, a warehousing robot state and warehousing environment information; then, based on the warehousing system environment map and a general task pool, constructing a task scheduling model; then, based on the task scheduling model, designing a task scheduling algorithm, and distributing tasks in a total task pool to each storage robot; then, aiming at the task pool of each storage robot, designing a task order optimization algorithm, and generating an execution plan of each storage robot; then, when the tasks in the general task pool are newly added or cancelled, the tasks in the general task pool are redistributed based on the current situation; and finally, for the dynamic fault condition of the storage robot, the scheduling center rolls back the task which is not completed by the fault storage robot to the scheduling center, and task allocation is carried out again.

Description

Technical Field

[0001] The present invention relates to the field of intelligent warehousing system control technology, and in particular to a task scheduling method for an intelligent warehousing system. Background Art

[0002] With the rapid development of e-commerce and logistics industries, intelligent warehousing systems have gradually become a key technology to improve warehouse management efficiency and reduce operating costs. Traditional warehousing systems mainly rely on the collaborative work of manual operations and mechanical equipment. Although they meet the needs of logistics and warehousing to a certain extent, they still have many shortcomings in terms of efficiency, accuracy and flexibility. Especially when the order processing volume is large and the variety is wide, traditional warehousing systems are prone to problems such as low efficiency, high error rate and high labor costs.

[0003] By introducing automation, information technology and intelligent technologies, the intelligent warehousing system can achieve efficient, accurate and flexible management of warehousing operations. The intelligent warehousing system can not only automatically complete tasks such as material storage, picking and handling, but also improve the utilization rate of warehousing resources and operating efficiency through data analysis and optimization algorithms. Among them, warehousing robots, as an important part of the intelligent warehousing system, play a key role. Warehouse robots can automatically navigate, avoid obstacles and transport goods in the warehouse, effectively reducing manpower requirements and operational errors, and improving overall operating efficiency.

[0004] However, the intelligent warehousing system faces a variety of complex dynamic events in actual operation, such as order changes, equipment failures, and environmental changes. These dynamic events place higher demands on the stability and efficiency of the system. In particular, when tasks change dynamically or warehouse robots fail, the system needs to respond quickly and reallocate tasks to ensure the continuity and efficiency of warehousing operations. Therefore, studying an intelligent warehousing system method that can handle dynamic events and optimize task scheduling is of great significance for improving the robustness and adaptability of the system.

[0005] In order to solve the above technical problems, the present invention proposes a task scheduling method for an intelligent warehousing system. Summary of the invention

[0006] The present invention provides a task scheduling method for an intelligent warehousing system, which is used to solve the above technical problems. First, based on the real-time data of order information, warehouse robot status and warehouse environment information, a warehouse system environment map and a total task pool are generated; then, based on the warehouse system environment map and the total task pool, a task scheduling model is constructed, the task scheduling model includes task set and robot set description, path constraints, resource constraints and task priority constraints, and takes minimizing path overhead, task completion time and robot energy consumption as the optimization goal; then, based on the task scheduling model, a task scheduling algorithm is designed to assign tasks in the total task pool to each warehouse robot to obtain a task pool for each warehouse robot; then, for each warehouse robot Based on the task pool of the robot, a task sequence optimization algorithm is designed to optimize the task execution order according to the distance between the task starting point and the end point and the task importance, and generate an execution plan for each warehouse robot; then, during the task execution process, the warehouse environment and task status are monitored in real time; when an abnormal situation is found, that is, new tasks are added and tasks are cancelled, the tasks in the total task pool are reallocated based on the current task execution status; finally, in the case of dynamic failure of the warehouse robot, the dispatch center rolls back the unfinished tasks of the faulty warehouse robot to the dispatch center, and then reallocates the tasks according to the current warehouse system environment map and the updated total task pool.

[0007] A task scheduling method for an intelligent warehousing system comprises the following steps:

[0008] S1. Generate a warehouse system environment map and a total task pool based on real-time data of order information, warehouse robot status, and warehouse environment information;

[0009] S2. Based on the warehouse system environment map and the total task pool, a task scheduling model is constructed. The task scheduling model includes a description of a task set and a robot set, path constraints, resource constraints, and task priority constraints, and minimizes path overhead, task completion time, and robot energy consumption as an optimization goal;

[0010] S3. Based on the task scheduling model, a task scheduling algorithm is designed to assign tasks in the total task pool to each warehouse robot to obtain a task pool for each warehouse robot;

[0011] S4. Design a task order optimization algorithm for each warehouse robot’s task pool, optimize the task execution order according to the distance between the task starting point and the end point and the task importance, and generate an execution plan for each warehouse robot;

[0012] S5. During the task execution process, the storage environment and task status are monitored in real time; when new tasks are added or cancelled in the total task pool, the tasks in the total task pool are reallocated based on the current task execution status;

[0013] S6. In the case of dynamic failure of the storage robot, the dispatch center rolls back the unfinished tasks of the faulty storage robot to the dispatch center, and then reallocates the tasks based on the current storage system environment map and the updated total task pool.

[0014] Furthermore, the warehouse system environment map and the total task pool are generated according to the real-time data of the order information, the warehouse robot status and the warehouse environment information; the grid method is selected to construct the warehouse system environment map, and the specific steps include:

[0015] A two-dimensional rectangular coordinate system of the storage environment is constructed, in which the x-axis represents the length of the storage, the y-axis represents the width of the storage, and the position of the equipment is represented by (x, y) coordinates; a grid of 1 unit length is set, and the grid width is set to be the same width as the passable road; the two-dimensional rectangular coordinate system includes: a picking platform 1, a road 2, a storage robot 3, a shelf 4 and a charging pile 5, the picking platform 1, the shelf 4 and the charging pile 5 are regarded as static obstacles, and the storage robot 3 is regarded as a dynamic obstacle; each storage robot moves in four directions on the road 2, up, down, left and right, and follows the traffic rule of driving on the right. The driving directions of adjacent passable roads are different, and the robots on the same road have the same driving direction. The distance between each storage robot is kept greater than 0.5; according to the above settings, the storage system environment map is established; at the same time, order information is obtained to build a total task pool for the intelligent warehousing system.

[0016] Furthermore, based on the warehouse system environment map and the total task pool, the task scheduling model is constructed. The task scheduling model includes a task set and a robot set, path constraints, resource constraints, and task priority constraints, and minimizes path overhead, task completion time, and robot energy consumption as the optimization goal. The task scheduling model establishment process is as follows:

[0017] Establish a task set S and a robot set R, where the task set S is a set of tasks that need to be assigned at time t, expressed as S = {s i |i=1,2,…,m}, where s i represents the task, i is an auxiliary variable; the robot set R is a set of storage robots that can be used to perform the task, expressed as R = {r l |l=1,2,…,n}, where r l represents the warehouse robot, l is an auxiliary variable;

[0018] Each task s in the task set S i The description form is in is the two-dimensional coordinate of the task starting point, is the two-dimensional coordinate of the task end point, t i Generate time for the task, h iThe importance of the task;

[0019] Each storage robot r in the robot set R l The description form is in is the current position coordinate of the storage robot, b l is the power of the storage robot, Z l It is the local task set of the warehouse robot;

[0020] Establish the local task set Z of the lth warehouse robot l , initialized to empty, each task is integrated into Z after task assignment l Middle; Z l There are multiple tasks in Z l ={s l1 ,s l2 ,…,s lk}, where s l1 ,s l2 ,…,s lk is the task assigned to the l-th storage robot, k is the number of tasks assigned to the l-th storage robot, and l is an auxiliary variable; D(r l ,Z l ) is the lth warehouse robot completing its local task set Z l The path cost is calculated as:

[0021] D(r l ,Z l )=L(s l1 )+L(s l2 )+...+L(s lk )+M(r l ,s l1 )+M(s l1 ,s l2 )+...+M(s l(k-1) ,s lk );

[0022] Among them, L(s lk ) is the lth storage robot completing a single task s lk The distance, M(s l(k-1) ,s lk ) is the lth storage robot from task s l(k-1) The end point moves to task s lk The transfer distance of the starting point;

[0023] The sum of the path costs of all robots completing their local task sets is represented by D all , the calculation formula is:

[0024]

[0025] The task execution optimization goal is to minimize the total cost of task execution. The specific optimization goals are as follows:

[0026] min(α*D all +β*T wei +γ*T max );

[0027]

[0028] Among them, α, β, and γ are weight parameters. is a decision variable, whose value is 0 or 1. Represents tasks j Assigned to storage robot l , Represents tasks j Not assigned to storage robot l ; Represents warehouse robot r l The number of tasks executed at time t, T over represents the sum of the delay times of all tasks, T wei represents the sum of weighted delay times of all tasks, T max Indicates the estimated end time of the system completing all tasks. The calculation formula is:

[0029]

[0030]

[0031]

[0032] in, For tasks l The time when the storage robot starts to execute, v l Represents warehouse robot r l speed.

[0033] Furthermore, based on the task scheduling model, a task scheduling algorithm is designed to assign tasks in the total task pool to each storage robot to obtain a task pool for each storage robot. The task scheduling algorithm includes the following steps:

[0034] S1. Initialization: Randomly generate an initial task scheduling plan;

[0035] S2. Neighborhood search: Generate a neighborhood solution of the current solution by exchanging the order of tasks or adjusting the order of task execution;

[0036] S3. Fitness evaluation: Calculate the fitness value of the current solution and the neighborhood solution. The fitness function is designed to minimize the total cost of task execution. The specific optimization objectives are as follows:

[0037] min(α*D all +β*T wei +γ*T max );

[0038]

[0039] S4. Acceptance criteria: Calculate the acceptance probability based on the fitness values ​​of the current solution and the neighboring solution, and decide whether to accept the neighboring solution. The higher the temperature, the greater the probability of accepting a worse solution, and the lower the temperature, the smaller the probability of accepting a worse solution.

[0040] S5. Temperature update: Use a linear decreasing strategy to gradually reduce the temperature;

[0041] S6. Termination condition: terminate when the maximum number of iterations is reached or the temperature drops to the preset value.

[0042] Furthermore, for each warehouse robot's task pool, a task order optimization algorithm is designed to optimize the task execution order according to the distance between the task starting point and the end point and the task importance, and generate an execution plan for each warehouse robot, including the following steps:

[0043] S10. Individual encoding: Encode each task sequence as an individual, each individual represents a task execution order, each individual consists of a set of integers, and each integer represents a task;

[0044] S20. Design fitness function: The goal of task sequence optimization is to find the optimal task sequence so that the weighted sum of the robot's travel distance to complete its local task list and the weighted overtime time of all tasks is minimized, that is,

[0045]

[0046] in, is the weight, f1 is the estimated shortest driving distance for the warehouse robot to execute the local task set, and f2 is the shortest task overdue time of the existing local task set of the warehouse robot. The calculation formulas of f1 and f2 are as follows:

[0047]

[0048] Among them, G(r l ,s lj ) represents the storage robot r l In the task list order from task s l1 To tasks lj The cumulative driving distance is calculated as follows:

[0049]

[0050] S30. Population setting: Initialize a population, including several individuals, each of which represents a possible task execution order. The population size is between 4N-6N, where N is the number of individual codes.

[0051] S40. Individual selection: Use the roulette wheel selection method to select individuals with high fitness from the population as parents;

[0052] S50. Crossover mutation: select two parent individuals, perform a crossover operation to generate new offspring individuals, and perform a mutation operation on the newly generated offspring individuals;

[0053] S60. Immune operation: select individuals with high fitness from the current population for cloning to generate multiple cloned individuals; Clone mutation: perform mutation operations on cloned individuals to maintain the diversity of solutions; Immune memory: save high-quality solutions in the immune pool to prevent forgetting the optimal solution;

[0054] S70. Termination condition: terminate when the maximum number of iterations is reached or the optimal value does not change.

[0055] Furthermore, during the task execution process, the storage environment and task status are monitored in real time; when new tasks are added or canceled in the total task pool, the tasks in the total task pool are reallocated based on the current task execution status, specifically:

[0056] S100. Adding new tasks: When new tasks are added to the total task pool, after the number of new tasks reaches a preset threshold, the task scheduling algorithm is used to uniformly allocate tasks; during this process, the system only recalculates and schedules tasks that have not yet been allocated, while the allocated tasks remain unchanged;

[0057] S200. Task cancellation: Delete the task according to the situation. If the task has not been assigned to the storage robot, delete the task directly; if the task has been assigned but has not started, delete the task and notify the corresponding storage robot to re-optimize the order of the remaining tasks; if the task has been assigned and is being executed, the storage robot needs to return the shelf to its original location and then execute the remaining tasks.

[0058] Furthermore, in the case of dynamic failure of the storage robot, the dispatch center rolls back the unfinished tasks of the faulty storage robot to the dispatch center, and then redistributes the tasks according to the current storage system environment map and the updated total task pool. The specific steps are as follows:

[0059] S1000. Task rollback: remove the unfinished tasks of the faulty robot from the local task set and roll them back to the task set;

[0060] S2000. Environment update: Update the environment map of the storage system to obtain the location and status change of the faulty robot;

[0061] S3000. Task reallocation: Using the task scheduling model and task scheduling algorithm, tasks are allocated to the updated total task pool. Compared with the prior art, the beneficial effects of the present invention are:

[0062] 1. The present invention realizes efficient task allocation and execution order optimization by designing task scheduling algorithm and task order optimization algorithm. The simulated annealing algorithm and particle swarm hybrid genetic algorithm are adopted to make task scheduling more flexible and accurate, and significantly improve the overall operation efficiency of the storage system.

[0063] 2. In response to dynamic events such as dynamic changes in tasks and failures of warehouse robots, the present invention proposes a corresponding rescheduling scheme, which enables it to quickly respond and adjust task allocation, effectively ensuring the continuity and stability of the system under complex working conditions, and improving the robustness and adaptability of the system.

[0064] 3. The present invention uses intelligent task scheduling and optimization algorithms to maximize the use of warehouse resources and reduce the total cost and time of task execution. In particular, the reallocation and optimization of unfinished tasks helps to improve the utilization rate and work efficiency of warehouse robots, reduce operating costs, and improve the overall performance of the system and customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a flow chart of a task scheduling method for an intelligent warehousing system;

[0066] Figure 2 It is the environment map of the warehouse system;

[0067] Figure 3 It is the flow chart of particle swarm hybrid genetic algorithm;

[0068] Figure 4 It is the process of handling dynamic changes in tasks. DETAILED DESCRIPTION

[0069] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] In modern logistics and supply chain management, task scheduling of intelligent warehousing systems has received increasing attention. Through advanced scheduling algorithms and optimization strategies, intelligent warehousing systems can effectively manage and arrange warehousing tasks, thereby improving the efficiency and accuracy of warehouse operations. However, in complex warehousing environments, such as multiple cargo locations, multiple orders, and dynamic demands, task scheduling systems face many challenges. Intelligent warehousing systems need to maintain stable and efficient scheduling performance under various working conditions, such as handling peak orders, dynamic task changes, and system interference. Traditional scheduling methods often find it difficult to ensure sufficient scheduling accuracy and response speed when faced with the nonlinear characteristics of the system and external interference, thereby affecting the efficiency and accuracy of warehousing operations. To this end, the present invention provides a task scheduling method for an intelligent warehousing system. In order to illustrate the feasibility of the present invention, it will be described below from the following specific embodiments.

[0071] Example

[0072] In the embodiment of the present application, a task scheduling method of an intelligent warehousing system is adopted, and the specific implementation process is as follows Figure 1 As shown, first, according to the real-time data of order information, warehouse robot status and warehouse environment information, the warehouse system environment map and the total task pool of the intelligent warehousing system are obtained, and a task scheduling model is constructed according to the warehouse system environment map and the total task pool; secondly, according to the task scheduling model, a task scheduling algorithm is designed to assign the tasks in the total task pool to each warehouse robot to obtain the task pool of each warehouse robot; then, according to the task pool of each warehouse robot, a task sequence optimization algorithm is designed to optimize the task execution order; finally, when an abnormality is found during the task execution process, the tasks in the total task pool are reallocated according to the current system status and task execution status.

[0073] The embodiment of the present application is task scheduling and optimization based on a grid map, and the specific implementation process is as follows:

[0074] Furthermore, based on the real-time data of order information, warehouse robot status and warehouse environment information, the warehouse system environment map and the total task pool are generated, wherein the grid method is selected to construct the warehouse system environment map, and the specific steps include:

[0075] Construct a two-dimensional rectangular coordinate system for the storage environment, where the x-axis represents the length of the storage, the y-axis represents the width of the storage, and the location of the equipment is represented by (x, y) coordinates; set a grid of 1 unit length, and set the grid width to the same as the width of the passable road; Figure 2As shown, the two-dimensional rectangular coordinate system includes: a picking platform 1, a road 2, a storage robot 3, a shelf 4 and a charging pile 5. The picking platform 1, the shelf 4 and the charging pile 5 are regarded as static obstacles, and the storage robot 3 is regarded as a dynamic obstacle; each storage robot moves in four directions on the road 2, up, down, left and right, and follows the traffic rule of driving on the right. The driving directions of adjacent passable roads are different, and the robots on the same road have the same driving direction. The distance between each storage robot is greater than 0.5; according to the above settings, the storage system environment map is established; at the same time, the order information is obtained to build the total task pool of the intelligent warehousing system.

[0076] In the embodiment of the present application, the technical solution of using the grid method to construct the warehouse system environment map can significantly improve the task scheduling accuracy and efficiency of the warehouse system. By clearly identifying obstacles and dynamic elements, the system can achieve real-time dynamic adjustment, optimize the task execution path, and effectively reduce collisions and traffic conflicts, thereby improving the operational safety of warehouse robots and the reliability of task execution.

[0077] A task scheduling model is constructed based on the warehouse system environment map and the total task pool. The task scheduling model establishment process is as follows:

[0078] Establish a task set S and a robot set R, where the task set S is a set of tasks that need to be assigned at time t, expressed as S = {s i |i=1,2,…,m}, each task s i It contains information such as the starting point, the end point, the generation time and the importance. The robot set R is a set of warehouse robots that can be used to perform tasks, expressed as R = {r i |i=1,2,…,n}, each robot r i Contains information such as current location, power level, and local task set;

[0079] Establish a task set S and a robot set R, where the task set S is a set of tasks that need to be assigned at time t, expressed as S = {s i |i=1,2,…,m}, where s i represents the task, i is an auxiliary variable; the robot set R is a set of storage robots that can be used to perform the task, expressed as R = {r l |l=1,2,…,n}, where r l represents the warehouse robot, l is an auxiliary variable;

[0080] Each task s in the task set S i The description form is in is the two-dimensional coordinate of the task starting point, is the two-dimensional coordinate of the task end point, t iGenerate time for the task, h i The importance of the task;

[0081] Each storage robot r in the robot set R l The description form is in is the current position coordinate of the storage robot, b l is the power of the storage robot, Z l It is the local task set of the warehouse robot;

[0082] Establish the local task set Z of the lth warehouse robot l , initialized to empty, each task is integrated into Z after task assignment l Middle; Z l There are multiple tasks in Z l ={s l1 ,s l2 ,…,s lk}, where s l1 ,s l2 ,…,s lk is the task assigned to the l-th storage robot, k is the number of tasks assigned to the l-th storage robot, and l is an auxiliary variable; D(r l ,Z l ) is the lth warehouse robot completing its local task set Z l The path cost is calculated as:

[0083] D(r l ,Z l )=L(s l1 )+L(s l2 )+...+L(s lk )+M(r l ,s l1 )+M(s l1 ,s l2 )+...+M(s l(k-1) ,s lk );

[0084] Among them, L(s lk ) is the lth storage robot completing a single task s lk The distance, M(s l(k-1) ,s lk ) is the lth storage robot from task s l(k-1) The end point moves to task s lk The transfer distance of the starting point;

[0085] The sum of the path costs of all robots completing their local task sets is represented by D all , the calculation formula is:

[0086]

[0087] The task execution optimization goal is to minimize the total cost of task execution. The specific optimization goals are as follows:

[0088] min(α*D all +β*T wei +γ*T max );

[0089]

[0090] Among them, α, β, and γ are weight parameters. is a decision variable, whose value is 0 or 1. Represents tasks j Assigned to storage robot l , Represents tasks j Not assigned to storage robot l ; Represents warehouse robot r l The number of tasks executed at time t, T over represents the sum of the delay times of all tasks, T wei represents the sum of weighted delay times of all tasks, T max Indicates the estimated end time of the system completing all tasks. The calculation formula is:

[0091]

[0092]

[0093] in, For tasks l The time when the storage robot starts to execute, v l Represents warehouse robot r l speed.

[0094] In the embodiment of the present application, the task scheduling model of the intelligent warehousing system is established with significant optimization effects. The model can improve the accuracy and efficiency of task allocation, reduce the delay time of task execution, and improve the overall operation efficiency and resource utilization of the warehousing system. In addition, the dynamic adjustment capability of the model helps to cope with changes in tasks and robot states, further enhance the adaptability and flexibility of the system, ensure that tasks are completed on time and maximize system performance.

[0095] Furthermore, according to the task scheduling model, a task scheduling algorithm is designed to assign tasks in the total task pool to each storage robot to obtain a task pool for each storage robot. The task scheduling algorithm includes the following steps:

[0096] S1. Initialization: Randomly generate an initial task scheduling plan;

[0097] S2. Neighborhood search: Generate a neighborhood solution of the current solution by exchanging the order of tasks or adjusting the order of task execution;

[0098] S3. Fitness evaluation: Calculate the fitness value of the current solution and the neighborhood solution. The fitness function is designed to minimize the total cost of task execution. The specific optimization objectives are as follows:

[0099] min(α*D all +β*T wei +γ*T max );

[0100]

[0101] S4. Acceptance criteria: Calculate the acceptance probability based on the fitness values ​​of the current solution and the neighboring solution, and decide whether to accept the neighboring solution. The higher the temperature, the greater the probability of accepting a worse solution, and the lower the temperature, the smaller the probability of accepting a worse solution.

[0102] S5. Temperature update: Use a linear decreasing strategy to gradually reduce the temperature;

[0103] S6. Termination condition: terminate when the maximum number of iterations is reached or the temperature drops to the preset value.

[0104] In the embodiment of the present application, the designed task scheduling algorithm can effectively optimize the task allocation process. The algorithm significantly improves the overall quality of the scheduling scheme through dynamic adjustment and evaluation methods, and reduces the total cost of task execution. By optimizing the acceptance criteria and temperature update strategy, the algorithm can avoid falling into the local optimal solution, thereby improving the global optimization ability of the system and the stability of the solution. These measures work together to improve the efficiency of task scheduling and the adaptability of the system.

[0105] Furthermore, according to the task pool of each storage robot, a task order optimization algorithm is designed to optimize the task execution order. The task order optimization algorithm is a particle swarm hybrid genetic algorithm. The algorithm flow chart is as follows: Figure 3 As shown, first, starting from the "Start" node, initialize the parameter settings and particle swarm. Next, set the time variable T = 1, and then enter the fitness calculation stage. In this stage, the algorithm calculates the fitness value based on the current state of the particle. Subsequently, update the individual optimal value and the group optimal value to find the best solution for each particle and the entire group respectively. Next, perform the optimization exchange of individuals and groups. After that, determine whether the end condition is met. If so, output the optimal result and end the algorithm; if not, increase the time variable T by 1 and continue with the next round of optimization. The whole process is repeated until the predetermined end condition is reached. The specific steps are:

[0106] S10. Individual encoding: Encode each task sequence as an individual, each individual represents a task execution order. Each individual consists of a set of integers, each integer represents a task.

[0107] S20. Design fitness function: The goal of task sequence optimization is to find the optimal task sequence so that the weighted sum of the robot's travel distance to complete its local task list and the weighted overtime time of all tasks is minimized, that is,

[0108]

[0109] in, is the weight, f1 is the estimated shortest driving distance of the local task set of the warehouse robot, and f2 is the shortest task overdue time of the existing local task set of the warehouse robot. The calculation formulas of f1 and f2 are as follows:

[0110]

[0111] Among them, G(r l ,s lj ) represents the storage robot r l In the task list order from task s l1 To tasks lj The cumulative driving distance is calculated as follows:

[0112]

[0113] S30. Population setting: Initialize a population, including several individuals, each of which represents a possible task execution order. The population size is between 4N-6N, where N is the number of individual codes.

[0114] S40. Individual selection: Use the roulette wheel selection method to select individuals with high fitness from the population as parents;

[0115] S50. Crossover mutation: select two parent individuals, perform a crossover operation to generate new offspring individuals, and perform a mutation operation on the newly generated offspring individuals;

[0116] S60. Termination condition: terminate when the maximum number of iterations is reached or the optimal value does not change.

[0117] In the embodiment of the present application, the task order optimization method based on the particle swarm hybrid genetic algorithm significantly improves the efficiency and accuracy of task scheduling. Through effective individual coding and fitness function design, the algorithm can optimize the order of task execution, reduce the driving distance and task overtime, thereby reducing the overall execution cost. In addition, population setting, individual selection and crossover mutation operations ensure global search capabilities and avoid the trouble of local optimal solutions. Ultimately, this method improves the adaptability and stability of the task scheduling system and enhances the overall operating efficiency of the warehouse robot.

[0118] For the situation where tasks change dynamically, online task allocation is used to handle dynamic changes in tasks. The specific processing process is as follows: Figure 4 The specific implementation process is as follows:

[0119] S100. Task addition: When new tasks are added to the total task pool, after the number of new tasks reaches a preset threshold, the task scheduling algorithm is used to uniformly allocate tasks; during this process, the system only recalculates and schedules tasks that have not yet been allocated, while the allocated tasks remain unchanged;

[0120] S200. Task reduction: Delete the task according to the situation. If the task has not been assigned to the storage robot, delete the task directly; if the task has been assigned but has not started to be executed, delete the task and notify the corresponding storage robot to re-optimize the order of the remaining tasks; if the task has been assigned and is being executed, the storage robot needs to return the shelf to its original location and then execute the remaining tasks.

[0121] In the embodiment of the present application, the online task allocation method is used to handle dynamically changing tasks, which significantly improves the flexibility and responsiveness of the system. The method can efficiently cope with the dynamic addition and deletion of tasks, ensuring the continuity of task scheduling and the operational stability of the system. By adjusting and optimizing task allocation in a timely manner, the system can minimize the interruption and delay of task execution, and improve the overall operational efficiency and resource utilization.

[0122] In the case of dynamic failure of the storage robot, the dispatch center rolls back the unfinished tasks of the faulty storage robot to the dispatch center, and then redistributes the tasks according to the current storage system environment map and the updated total task pool. The specific steps are as follows:

[0123] S1000. Task rollback: remove the unfinished tasks of the faulty robot from the local task set and roll them back to the total task pool;

[0124] S2000. Environment update: Update the environment map of the storage system to obtain the location and status change of the faulty robot;

[0125] S3000. Task reallocation: Use the task scheduling model and task scheduling algorithm to allocate tasks to the updated total task pool.

[0126] In the embodiment of the present application, the method for handling dynamic failures of storage robots significantly enhances the robustness and reliability of the system. Through effective rollback and priority sorting, combined with the simulated annealing algorithm to reallocate tasks, it is possible to efficiently deal with robot failures and reduce interruptions and delays in task execution. This mechanism improves the system's adaptability to sudden failures, ensuring the smooth completion of tasks and the continued stable operation of the system.

[0127] The comparison table of the results of the traditional particle swarm optimization algorithm and the particle swarm hybrid genetic optimization algorithm is as follows. Table 1 is the result of the particle swarm optimization algorithm, and Table 2 is the result of the particle swarm hybrid genetic optimization algorithm.

[0128] Table 1 Results of particle swarm optimization algorithm

[0129]

[0130]

[0131] Table 2 Particle swarm hybrid genetic optimization algorithm

[0132]

[0133] Compared with the particle swarm optimization algorithm, the particle swarm hybrid genetic optimization algorithm shows better performance under different numbers of robot configurations. Specifically, whether it is 20, 40 or 60 robots, the hybrid genetic algorithm can achieve a higher average AGV driving distance, while reducing the total time to complete the task and the average time to complete the task. This shows that by combining the genetic algorithm, the particle swarm optimization algorithm has been significantly improved in path planning and task scheduling, making the robot more efficient and completing tasks faster.

[0134] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A task scheduling method for an intelligent warehousing system, characterized in that: The following steps are involved: S1. Generate a warehouse system environment map and a total task pool based on real-time data of order information, warehouse robot status, and warehouse environment information; S2. Based on the warehouse system environment map and the total task pool, a task scheduling model is constructed. The task scheduling model includes a description of a task set and a robot set, path constraints, resource constraints, and task priority constraints, and minimizes path overhead, task completion time, and robot energy consumption as an optimization goal; S3. Based on the task scheduling model, a task scheduling algorithm is designed to assign tasks in the total task pool to each warehouse robot to obtain a task pool for each warehouse robot; S4. Design a task order optimization algorithm for each warehouse robot’s task pool, optimize the task execution order according to the distance between the task starting point and the end point and the task importance, and generate an execution plan for each warehouse robot; S5. During the task execution process, the storage environment and task status are monitored in real time; when new tasks are added or cancelled in the total task pool, the tasks in the total task pool are reallocated based on the current task execution status; S6. In the case of dynamic failure of the storage robot, the dispatch center rolls back the unfinished tasks of the faulty storage robot to the dispatch center, and then reallocates the tasks based on the current storage system environment map and the updated total task pool.

2. The task scheduling method of the intelligent warehousing system according to claim 1 is characterized in that: Generate the warehouse system environment map and the total task pool according to the real-time data of order information, warehouse robot status and warehouse environment information; The grid method is selected to construct the storage system environment map, and the specific steps include: A two-dimensional rectangular coordinate system of a storage environment is constructed, wherein the x-axis represents the length of the storage, the y-axis represents the width of the storage, and the position of the equipment is represented by (x, y) coordinates; a grid of 1 unit length is set, and the grid width is set to be the same as the width of a passable road; the two-dimensional rectangular coordinate system includes: a picking platform (1), a road (2), a storage robot (3), a shelf (4) and a charging pile (5), wherein the picking platform (1), the shelf (4) and the charging pile (5) are regarded as static obstacles, and the storage robot (3) is regarded as a dynamic obstacle; each storage robot moves in four directions, up, down, left and right, on the road (2), following the traffic rule of driving on the right, with adjacent passable roads driving in different directions, and robots on the same road driving in the same direction, and the distance between each storage robot is kept greater than 0.5; based on the above settings, an environmental map of the storage system is established; and order information is obtained at the same time to construct a total task pool of the intelligent storage system.

3. The task scheduling method of the intelligent warehousing system according to claim 1 is characterized in that: Based on the warehouse system environment map and the total task pool, the task scheduling model is constructed. The task scheduling model includes a task set and a robot set, path constraints, resource constraints, and task priority constraints, and minimizes path overhead, task completion time, and robot energy consumption as the optimization goal. The task scheduling model establishment process is as follows: Establish a task set S and a robot set R, where the task set S is a set of tasks that need to be assigned at time t, expressed as S = {s i |i=1,2,…,m}, where s i represents the task, i is an auxiliary variable; the robot set R is a set of warehouse robots that can be used to perform tasks, expressed as R = {r l |l=1,2,…,n}, where r l represents the storage robot, l is an auxiliary variable; Each task s in the task set S i The description form is in is the two-dimensional coordinate of the starting point of the task, is the two-dimensional coordinate of the task end point, t i The task generation time, h i The importance of the task; Each storage robot r in the robot set R l The description form is in is the current position coordinate of the storage robot, b l is the power of the storage robot, Z l It is the local task set of the warehouse robot; Establish the local task set Z of the lth warehouse robot l , initialized to empty, each task is integrated into Z after task assignment l Middle; Z l There are multiple tasks in Z l ={s l1 ,s l2 ,…,s lk }, where s l1 ,s l2 ,…,s lk is the task assigned to the l-th storage robot, k is the number of tasks assigned to the l-th storage robot; D(r l ,Z l ) is the lth warehouse robot completing its local task set Z l The path cost is calculated as: D(r l ,Z l )=L(s l1 )+L(s l2 )+...+L(s lk )+M(r l ,s l1 )+M(s l1 ,s l2 )+... +M(s l(k-1) ,s lk ); Among them, L(s lk ) is the lth storage robot completing a single task s lk The distance, M(s l(k-1) ,s lk ) is the lth storage robot from task s l(k-1) The end point moves to task s lk The transfer distance of the starting point; The sum of the path costs of all warehouse robots to complete their local task sets is expressed as D all , the calculation formula is: The task execution optimization goal is to minimize the total cost of task execution. The specific optimization goals are as follows: min(α*D all +β*T wei +γ*T max ); Among them, α, β, and γ are weight parameters. is a decision variable, whose value is 0 or 1. Represents tasks j Assigned to storage robot l , Represents tasks j Not assigned to storage robot l ; Represents warehouse robot r l The number of tasks executed at time t, T over represents the sum of the delay times of all tasks, T wei represents the sum of weighted delay times of all tasks, T max Indicates the estimated end time of the system completing all tasks. The calculation formula is: in, For tasks l The time when the storage robot starts to execute, v l Represents warehouse robot r l speed.

4. The task scheduling method of the intelligent warehousing system according to claim 1 is characterized in that: Based on the task scheduling model, a task scheduling algorithm is designed to assign tasks in the total task pool to each warehouse robot to obtain a task pool for each warehouse robot. The task scheduling algorithm includes the following steps: S1. Initialization: Randomly generate an initial task scheduling plan; S2. Neighborhood search: Generate a neighborhood solution of the current solution by exchanging the order of tasks or adjusting the order of task execution; S3. Fitness evaluation: Calculate the fitness value of the current solution and the neighborhood solution. The fitness function is designed to minimize the total cost of task execution. The specific optimization objectives are as follows: min(α*D all +β*T wei +γ*T max ); S4. Acceptance criteria: Calculate the acceptance probability based on the fitness values ​​of the current solution and the neighboring solution and the current temperature, and decide whether to accept the neighboring solution. The higher the temperature, the greater the probability of accepting a worse solution, and the lower the temperature, the smaller the probability of accepting a worse solution. S5. Temperature update: Use a linear decreasing strategy to gradually reduce the temperature; S6. Termination condition: terminate when the maximum number of iterations is reached or the temperature drops to the preset value.

5. The task scheduling method of the intelligent warehousing system according to claim 4 is characterized in that: For each warehouse robot's task pool, a task order optimization algorithm is designed to optimize the task execution order according to the distance between the task starting point and the end point and the task importance, and generate an execution plan for each warehouse robot, including the following steps: S10. Individual encoding: Encode each task sequence as an individual, each individual represents a task execution order, each individual consists of a set of integers, and each integer represents a task; S20. Design fitness function: The goal of task sequence optimization is to find the optimal task sequence so that the weighted sum of the robot's travel distance to complete its local task list and the weighted overtime time of all tasks is minimized, that is, in, is the weight, f1 is the estimated shortest driving distance for the warehouse robot to execute the local task set, and f2 is the shortest task overdue time of the existing local task set of the warehouse robot. The calculation formulas of f1 and f2 are as follows: Among them, G(r l ,s lj ) represents the storage robot r l In the task list order from task s l1 To tasks lj The cumulative driving distance is calculated as follows: S30. Population setting: Initialize a population, including several individuals, each of which represents a possible task execution order. The population size is between 4N-6N, where N is the number of individual codes. S40. Individual selection: Use the roulette wheel selection method to select individuals with high fitness from the population as parents; S50. Crossover mutation: select two parent individuals, perform a crossover operation to generate new offspring individuals, and perform a mutation operation on the newly generated offspring individuals; S60. Immune operation: select individuals with high fitness from the current population for cloning to generate multiple cloned individuals; Clone mutation: perform mutation operations on cloned individuals to maintain the diversity of solutions; Immune memory: save high-quality solutions in the immune pool to prevent forgetting the optimal solution; S70. Termination condition: terminate when the maximum number of iterations is reached or the optimal value does not change.

6. The task scheduling method of the intelligent warehousing system according to claim 1 is characterized in that: During the task execution process, the storage environment and task status are monitored in real time; when new tasks are added or canceled in the total task pool, the tasks in the total task pool are reallocated based on the current task execution status, specifically: S100. Adding new tasks: When new tasks are added to the total task pool, after the number of new tasks reaches a preset threshold, the task scheduling algorithm is used to uniformly allocate tasks; during this process, the system only recalculates and schedules tasks that have not yet been allocated, while the allocated tasks remain unchanged; S200. Task cancellation: Delete the task according to the situation. If the task has not been assigned to the storage robot, delete the task directly; if the task has been assigned but has not started, delete the task and notify the corresponding storage robot to re-optimize the order of the remaining tasks; if the task has been assigned and is being executed, the storage robot needs to return the shelf to its original location and then execute the remaining tasks.

7. The task scheduling method of the intelligent warehousing system according to claim 6 is characterized in that: In the case of dynamic failure of the storage robot, the dispatch center rolls back the unfinished tasks of the faulty storage robot to the dispatch center, and then redistributes the tasks according to the current storage system environment map and the updated total task pool. The specific steps are as follows: S1000. Task rollback: remove the unfinished tasks of the faulty robot from the local task set and roll them back to the task set; S2000. Environment update: Update the environment map of the storage system to obtain the location and status change of the faulty robot; S3000. Task reallocation: Use the task scheduling model and task scheduling algorithm to allocate tasks to the updated total task pool.

Images