Method, device and system for optimizing whole process of in-library operation based on DMPC improved by simulated annealing algorithm

Abstract

The invention relates to the technical field of intelligent warehousing, in particular to an in-library operation whole-process optimization method, device and system for improving DMPC based on a simulated annealing algorithm. The method comprises the following steps: firstly, constructing a mathematical model of a whole process of in-library operation, wherein the mathematical model covers a plurality of links such as storage location allocation, path planning and task scheduling; then generating a preliminary optimization scheme by utilizing dynamic matrix predictive control (DMPC); secondly, improving an optimization result of the DMPC through a simulated annealing algorithm, avoiding a local optimal solution, and improving a global optimization effect; and finally, the working state of each operation unit in the library is dynamically adjusted according to an optimization result, so that balanced utilization of resources and maximization of operation efficiency are realized. According to the method, the DMPC is improved by introducing the simulated annealing algorithm, global optimization of the whole process of in-library operation is achieved, the resource utilization rate and the operation efficiency are improved, and the method can adapt to dynamic order demand changes.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent warehousing, and specifically relates to a method, device and system for optimizing the entire process of in-warehouse operations based on improved DMPC using a simulated annealing algorithm. Background Art

[0002] With the rapid development of the logistics industry, full-process optimization methods, devices and systems for warehouse operations have been widely used in warehouse management and operation optimization. However, there are still many problems with the existing technologies in practical applications. For example, existing methods mainly focus on warehouse layout optimization and picking path planning, and often ignore the comprehensive collaborative optimization of the entire operation process, resulting in inefficiency and insufficient resource utilization in the case of large-scale warehouses and dynamic order demand changes.

[0003] After searching, the patent with publication number CN113034084B discloses a method and terminal for dynamic configuration of unitized intelligent warehouses. This technology forms a unitized warehouse through virtualization technology, including manual fulfillment units, automated fulfillment units, grid fulfillment units and robot fulfillment units, and assembles orders and packages items through order merging units to reduce cross-unit cross-operation and walking distance in the warehouse, thereby reducing the total cost of order fulfillment. However, this technology has the following shortcomings: First, there is a lack of global optimization, and the entire warehouse operation process has not been optimized from a global perspective. The connection and coordination between various links need to be improved; second, the static matching method is difficult to adapt to the rapid changes and dynamic adjustments of order requirements, and the system flexibility and response speed are insufficient; finally, the resource utilization problem has not been effectively solved, and how to dynamically adjust the workload of each fulfillment unit to achieve balanced resource utilization is still a challenge.

[0004] In order to improve performance, some manufacturers have tried to improve the level of automation by introducing more advanced scheduling algorithms and intelligent equipment. However, such improvements often face the dual challenges of high algorithm complexity, difficulty in implementation, and difficulty in achieving stability and efficiency in actual applications.

[0005] The above problems show that the existing traditional warehouse management technology is difficult to effectively cope with the new requirements for precise control and rapid response under complex working conditions. Therefore, the present invention provides a method, device and system for optimizing the whole process of warehouse operations based on DMPC (dynamic matrix predictive control) improved by simulated annealing algorithm to overcome these shortcomings and provide a new solution that is more intelligent, efficient and adaptable to changing environments.

[0006] Specifically, the present invention improves DMPC through simulated annealing algorithm, and can realize dynamic resource allocation and operation optimization in multiple fulfillment units. This method can not only reduce cross-unit cross-operation and walking distance in the warehouse, but also quickly adjust the working status of each fulfillment unit when facing order demand fluctuations, ensuring full utilization of resources and optimization of operation efficiency. In addition, through global optimization and dynamic adjustment, the present invention can better adapt to different types of fulfillment scenarios and improve the stability and reliability of the overall system. Summary of the invention

[0007] The present invention relates to the field of intelligent warehousing technology, and specifically to a method, device and system for optimizing the entire process of in-warehouse operations based on improved DMPC using a simulated annealing algorithm. The present invention aims to solve the problems of lack of global optimization of in-warehouse operation processes, difficulty of static matching in adapting to dynamic demand changes and low resource utilization in the prior art, and improves dynamic matrix predictive control (DMPC) by introducing a simulated annealing algorithm to achieve intelligent optimization of the entire process of in-warehouse operations.

[0008] In the first aspect, the present invention provides a method for optimizing the whole process of in-warehouse operations based on the improved DMPC using the simulated annealing algorithm. The method comprises the following steps: first, constructing a mathematical model of the whole process of in-warehouse operations, covering multiple links such as storage location allocation, path planning, and task scheduling; second, using dynamic matrix predictive control (DMPC) to predict and control the in-warehouse operation process and generate a preliminary optimization plan; then, introducing the simulated annealing algorithm to further improve the optimization results of DMPC, and avoiding local optimal solutions and improving the optimization effect through the global search capability of the simulated annealing algorithm; finally, dynamically adjusting the working status of each operation unit in the warehouse according to the optimization results to achieve balanced resource utilization and maximized operation efficiency.

[0009] In the second aspect, the present invention provides a device for optimizing the whole process of in-warehouse operations based on the improved DMPC by the simulated annealing algorithm. The device includes: a model building module for building a mathematical model of the whole process of in-warehouse operations; a DMPC optimization module for generating a preliminary optimization scheme based on dynamic matrix predictive control; a simulated annealing improvement module for improving the optimization results of DMPC by the simulated annealing algorithm; and a dynamic adjustment module for dynamically adjusting the working status of each operation unit in the warehouse according to the optimization results.

[0010] In the third aspect, the present invention provides a system for optimizing the whole process of in-library operations based on the simulated annealing algorithm to improve DMPC. The system comprises: one or more processors; a memory; one or more computer programs; wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and when the one or more computer programs are executed by the processors, the above method is implemented.

[0011] Preferably, the mathematical model of the entire process of in-warehouse operations includes a warehouse location allocation model, a path planning model and a task scheduling model, which are respectively used to optimize warehouse location utilization, reduce path intersections and improve task execution efficiency.

[0012] Preferably, the dynamic matrix predictive control (DMPC) generates a preliminary optimization plan by predicting the future state of the operations in the warehouse, thereby ensuring the real-time and dynamic adaptability of the optimization plan.

[0013] Preferably, the simulated annealing algorithm gradually reduces the search range by introducing temperature parameters and annealing strategies, avoids falling into local optimal solutions, and improves the global optimization effect.

[0014] Preferably, the dynamic adjustment module adjusts the working status of each operating unit in the library in real time according to the optimization result to ensure balanced utilization of resources and maximization of operating efficiency.

[0015] Preferably, the model building module performs hierarchical modeling of the operation process in the library through hierarchical analysis method to ensure the comprehensiveness and accuracy of the model.

[0016] Preferably, the DMPC optimization module generates a preliminary optimization solution by introducing a multi-objective optimization function and comprehensively considering the storage space utilization, path length and task execution time.

[0017] Preferably, the simulated annealing improvement module introduces an adaptive annealing strategy to dynamically adjust temperature parameters according to feedback of optimization results to improve optimization efficiency.

[0018] Preferably, the dynamic adjustment module introduces an intelligent scheduling algorithm to dynamically adjust the working status of each operating unit in the library according to the optimization results, thereby ensuring balanced utilization of resources and maximization of operating efficiency.

[0019] Operation principle: This invention constructs a mathematical model of the entire process of the warehouse operation, uses dynamic matrix predictive control (DMPC) to generate a preliminary optimization plan, and improves the optimization results through simulated annealing algorithm to avoid local optimal solutions and improve the global optimization effect. Finally, the working status of each operation unit in the warehouse is dynamically adjusted according to the optimization results to achieve balanced resource utilization and maximized operation efficiency.

[0020] Beneficial effects: The present invention improves DMPC by introducing a simulated annealing algorithm, thereby achieving global optimization of the entire process of in-warehouse operations, improving resource utilization and operating efficiency, being able to adapt to changes in dynamic order demand, and solving the deficiencies in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1It is a flowchart of a method for optimizing the entire process of in-library operations based on a simulated annealing algorithm to improve DMPC provided in an embodiment of the present application.

[0022] Figure 2 It is a schematic diagram of the composition of a device for optimizing the entire process of in-store operations based on improved DMPC using a simulated annealing algorithm, provided in an embodiment of the present application.

[0023] Figure 3 It is a schematic diagram of the composition of a full-process optimization system for in-store operations based on improved DMPC using a simulated annealing algorithm provided in an embodiment of the present application.

[0024] Figure 4 It is a schematic diagram of constructing a mathematical model of the entire process of in-warehouse operations provided in an embodiment of the present application.

[0025] Figure 5 It is a workflow diagram of the dynamic matrix predictive control (DMPC) optimization module provided in the embodiment of the present application.

[0026] Figure 6 It is a workflow diagram of the simulated annealing algorithm improvement module provided in the embodiment of the present application.

[0027] Figure 7 It is a workflow diagram of the dynamic adjustment module provided in the embodiment of the present application.

[0028] Figure 8 It is a schematic diagram of the operation principle of the full-process optimization system for in-warehouse operations provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method part of the embodiments of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0030] The embodiment of the present application provides a method, device and system for optimizing the whole process of in-store operations based on the simulated annealing algorithm to improve DMPC. The execution subject of the method is a computer system, including but not limited to a server, a personal computer, a laptop computer, a tablet computer, a smart phone, etc. The embodiment of the present application provides a method for optimizing the whole process of in-store operations based on the simulated annealing algorithm to improve DMPC, such as Figure 1 As shown, the method includes:

[0031] Step S100: Construct a mathematical model of the entire process of in-warehouse operations. This step first requires the collection of various data on in-warehouse operations, including data on warehouse location allocation, path planning, task scheduling, and other links. These data can come from various channels such as historical records, sensor data, and manual records. For example, warehouse location allocation data includes the current warehouse location status (free, occupied, under maintenance, etc.), warehouse location capacity, type of stored and retrieved goods, etc.; path planning data includes all path nodes in the warehouse, the distance between each node, possible obstacles, etc.; task scheduling data includes the current task list, the priority of each task, the execution time and location of the task, etc.

[0032] Using these data, a mathematical model of the entire process of warehouse operations is constructed, covering the warehouse location allocation model, path planning model and task scheduling model. The warehouse location allocation model is used to optimize the utilization of warehouse locations and reduce the waiting time for goods storage and retrieval; the path planning model is used to reduce path intersections and improve path utilization; the task scheduling model is used to improve task execution efficiency and reduce conflicts between tasks. The specific construction method of the mathematical model can adopt the hierarchical analysis method and hierarchical modeling to ensure the comprehensiveness and accuracy of the model. For example, the warehouse operation process can be divided into multiple levels, each level handles different subtasks respectively, and the model of each layer is gradually optimized through the optimization algorithm, and finally a full-process optimization model of the entire warehouse operation is formed.

[0033] Step S200: Generate a preliminary optimization plan using dynamic matrix predictive control (DMPC). In this step, it is first necessary to design a dynamic matrix predictive control (DMPC) algorithm to generate a preliminary optimization plan by predicting the future state of the operations in the warehouse. The core of the DMPC algorithm is to predict possible changes in operations in the future based on the current operation status, and generate an optimization plan based on these prediction results. For example, assuming that there are currently multiple tasks to be processed, the DMPC algorithm will predict the execution order, path selection, warehouse location allocation and other information of each task in the future based on the current task list, task priority, warehouse location status, path information and other data, and generate a preliminary optimization plan.

[0034] The prediction process of the DMPC algorithm can use a state space model to represent each state of the job in the library as a state vector, and predict the future state through the state transfer equation. For example, assuming that the current state vector is x(t) and the future state vector is x(t+1), the state transfer equation can be expressed as x(t+1)=Ax(t)+Bu(t), where A is the state transfer matrix, B is the control input matrix, and u(t) is the current control input (such as task execution order, path selection, etc.). By solving the state transfer equation, the future state can be predicted and a preliminary optimization plan can be generated.

[0035] Step S300: Introduce simulated annealing algorithm to improve the optimization result of DMPC. After generating the preliminary optimization scheme, the optimization result of DMPC is further improved by introducing simulated annealing algorithm. Simulated annealing algorithm is a global search algorithm. By simulating the physical annealing process, the search range is gradually reduced to avoid falling into the local optimal solution and improve the global optimization effect. The specific steps include:

[0036] First, initialize the parameters of the simulated annealing algorithm, including the initial temperature T0, the termination temperature Tend, the cooling coefficient α, etc. The initial temperature T0 is set high and gradually reduced until the termination temperature Tend is reached. The cooling coefficient α is usually set between 0.8 and 0.99, which determines the magnitude of each cooling. Then, a candidate solution is generated based on the current optimization solution, and the fitness value of the candidate solution (such as task completion time, path length, storage location utilization, etc.) is calculated. If the fitness value of the candidate solution is better than the current solution, the candidate solution is accepted; otherwise, the candidate solution is accepted with a certain probability, and the probability calculation formula is P = exp(-(f(new)-f(old)) / T), where f(new) and f(old) are the fitness values ​​of the candidate solution and the current solution, respectively, and T is the current temperature. Through this random acceptance mechanism, the simulated annealing algorithm can jump out of the local optimal solution during the search process and explore a wider solution space.

[0037] Step S400: Dynamically adjust the working status of each operating unit in the library according to the optimization results. After generating the final optimization plan, dynamically adjust the working status of each operating unit in the library according to the optimization results to ensure balanced utilization of resources and maximized operating efficiency. The specific steps include:

[0038] First, control instructions are generated according to the optimization plan and sent to each operation unit. For example, if the priority of a task in the optimization plan is increased, a corresponding control instruction is generated to instruct the task to be executed first. Secondly, the execution status of each operation unit is monitored in real time, and the optimization plan is dynamically adjusted according to the feedback results. For example, if an abnormality occurs during the execution of a task, such as equipment failure, insufficient materials, etc., the execution order of the tasks is adjusted in time or resources are reallocated. Finally, through intelligent scheduling algorithms, such as genetic algorithms and ant colony algorithms, the allocation and scheduling of resources are optimized to ensure balanced utilization of resources and maximized operation efficiency.

[0039] As an implementation method, the mathematical model of the entire process of the in-library operation constructed in step S100 can be further refined into the following sub-steps:

[0040] Step S101: Collect various data of operations in the warehouse. Various data of operations in the warehouse are collected through channels such as sensors, historical records, and manual records, including warehouse location status, task lists, and path information.

[0041] Step S102: Use the analytic hierarchy process to model the warehouse operation process in layers. According to the different links of the warehouse operation, the entire process is divided into multiple layers, and each layer handles different subtasks to ensure the comprehensiveness and accuracy of the model. For example, the warehouse operation process can be divided into the warehouse location allocation layer, the path planning layer, and the task scheduling layer. The model of each layer is gradually optimized through the optimization algorithm, and finally a full-process optimization model of the entire warehouse operation is formed.

[0042] As an implementation method, the generation of a preliminary optimization solution using dynamic matrix predictive control in step S200 can be further refined into the following sub-steps:

[0043] Step S201: Design a dynamic matrix predictive control algorithm. According to the data of the operations in the warehouse, a DMPC algorithm is designed to generate a preliminary optimization plan by predicting the future state of the operations in the warehouse.

[0044] Step S202: Construct a state space model. Represent each state of the operation in the library as a state vector, and predict the future state through the state transfer equation. For example, assuming that the current state vector is x(t) and the future state vector is x(t+1), the state transfer equation can be expressed as x(t+1)=Ax(t)+Bu(t), where A is the state transfer matrix, B is the control input matrix, and u(t) is the current control input.

[0045] Step S203: Solve the state transition equation and generate a preliminary optimization plan. By solving the state transition equation, the future state is predicted and a preliminary optimization plan is generated. For example, assuming that there are multiple tasks to be processed, the DMPC algorithm will predict the execution order, path selection, storage location allocation and other information of each task in the future period of time based on the current task list, task priority, storage location status, path information and other data, and generate a preliminary optimization plan.

[0046] As an implementation method, the introduction of the simulated annealing algorithm in step S300 to improve the optimization result of DMPC can be further refined into the following sub-steps:

[0047] Step S301: Initialize the parameters of the simulated annealing algorithm, including the initial temperature T0, the end temperature Tend, the cooling coefficient α, etc. The initial temperature T0 is set high and gradually reduced until the end temperature Tend is reached. The cooling coefficient α is usually set between 0.8 and 0.99, which determines the extent of each cooling.

[0048] Step S302: Generate candidate solutions and calculate fitness values. Generate a candidate solution based on the current optimization solution and calculate the fitness value of the candidate solution. If the fitness value of the candidate solution is better than the current solution, accept the candidate solution; otherwise, accept the candidate solution with a certain probability. The probability calculation formula is P = exp(-(f(new)-f(old)) / T), where f(new) and f(old) are the fitness values ​​of the candidate solution and the current solution respectively, and T is the current temperature.

[0049] Step S303: gradually lower the temperature and repeat the above process. By gradually lowering the temperature, the search range is gradually reduced to avoid falling into a local optimal solution and improve the global optimization effect.

[0050] As an implementation method, dynamically adjusting the working status of each operation unit in the library according to the optimization result in step S400 can be further refined into the following sub-steps:

[0051] Step S401: Generate control instructions and send them to each operating unit. Generate control instructions according to the optimization plan and send them to each operating unit to instruct it to perform corresponding tasks.

[0052] Step S402: Monitor the execution status of each operation unit in real time. Monitor the execution status of each operation unit in real time through sensors, monitoring systems, etc., and dynamically adjust the optimization plan according to the feedback results.

[0053] Step S403: Optimize resource allocation and scheduling through intelligent scheduling algorithms, such as genetic algorithms, ant colony algorithms, etc., to optimize resource allocation and scheduling to ensure balanced resource utilization and maximized operating efficiency.

[0054] Through the above implementation mode, the present invention provides a method for optimizing the entire process of in-warehouse operations based on the improved DMPC using the simulated annealing algorithm, which can effectively solve the problems in the prior art of lack of global optimization of in-warehouse operation processes, difficulty of static matching in adapting to dynamic demand changes, and low resource utilization, improve resource utilization and operation efficiency, adapt to dynamic order demand changes, and ensure efficient, flexible and intelligent in-warehouse operations.

Claims

1. A method for optimizing the whole process of in-store operations based on the simulated annealing algorithm to improve DMPC, characterized in that: The method comprises: Constructing a mathematical model of the entire process of in-warehouse operations, the mathematical model includes a warehouse location allocation model, a path planning model, and a task scheduling model; Use dynamic matrix predictive control (DMPC) to predict and control the operation process within the warehouse and generate a preliminary optimization plan; The simulated annealing algorithm is introduced to improve the optimization result of DMPC, and the preliminary optimization scheme is optimized through the global search capability of the simulated annealing algorithm; Dynamically adjust the working status of each operating unit in the library according to the optimization results.

2. The method for optimizing the whole process of in-store operations based on the simulated annealing algorithm to improve DMPC as claimed in claim 1, characterized in that: The mathematical model for constructing the entire process of the in-library operation includes: Collect various data of operations in the warehouse, including warehouse location status, task list and path information; The hierarchical analysis method is used to perform hierarchical modeling on the operation process in the warehouse, and the hierarchical modeling includes a warehouse location allocation layer, a path planning layer, and a task scheduling layer.

3. The method for optimizing the whole process of in-store operations based on improved DMPC by simulated annealing algorithm as claimed in claim 1, characterized in that: The method of generating a preliminary optimization scheme using dynamic matrix predictive control (DMPC) includes: Design a dynamic matrix predictive control algorithm to generate a preliminary optimization plan by predicting the future state of operations in the warehouse; Construct a state space model, represent each state of the operation in the library as a state vector, and predict the future state through the state transition equation; Solve the state transition equation and generate a preliminary optimization solution.

4. The method for optimizing the whole process of in-store operations based on improved DMPC by simulated annealing algorithm as claimed in claim 1, characterized in that: The method of introducing the simulated annealing algorithm to improve the optimization result of DMPC includes: Initializing parameters of the simulated annealing algorithm, the parameters including initial temperature, end temperature and temperature reduction coefficient; Generate candidate solutions and calculate fitness values, and decide whether to accept the candidate solutions according to the fitness values; Gradually lower the temperature, repeatedly generate candidate solutions and calculate fitness values ​​until the termination temperature is reached.

5. The method for optimizing the whole process of in-store operations based on the simulated annealing algorithm to improve DMPC as claimed in claim 1, characterized in that: The dynamically adjusting the working status of each operating unit in the library according to the optimization result includes: Generate control instructions and send them to each operating unit to instruct it to perform the corresponding tasks; Monitor the execution status of each operation unit in real time and dynamically adjust the optimization plan based on the feedback results; Optimize resource allocation and scheduling through intelligent scheduling algorithms.

6. A device for optimizing the whole process of in-store operations based on the simulated annealing algorithm to improve DMPC, characterized in that: The device comprises: Model building module, used to build a mathematical model of the entire process of operations in the library; DMPC optimization module, used to generate preliminary optimization solutions using dynamic matrix predictive control (DMPC); The simulated annealing improvement module is used to improve the optimization results of DMPC through the simulated annealing algorithm; The dynamic adjustment module is used to dynamically adjust the working status of each operation unit in the library according to the optimization results.

7. The device for optimizing the whole process of in-store operations based on improved DMPC using simulated annealing algorithm as claimed in claim 6, characterized in that: The model building module performs hierarchical modeling on the in-warehouse operation process through the hierarchical analysis method, and the hierarchical modeling includes a warehouse location allocation layer, a path planning layer and a task scheduling layer.

8. The device for optimizing the whole process of in-store operations based on improved DMPC using simulated annealing algorithm as claimed in claim 6, characterized in that: The DMPC optimization module constructs a state space model, represents each state of the operation in the library as a state vector, and predicts the future state through the state transition equation.

9. The device for optimizing the whole process of in-store operations based on improved DMPC using simulated annealing algorithm as claimed in claim 6, characterized in that: The simulated annealing improvement module generates candidate solutions and calculates fitness values ​​by initializing parameters of the simulated annealing algorithm, and gradually reduces the temperature to optimize the preliminary optimization solution.

10. A full-process optimization system for in-warehouse operations based on improved DMPC using simulated annealing algorithm, characterized in that: The system comprises: one or more processors; Memory; One or more computer programs; wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and when the one or more computer programs are executed by the processors, the method according to any one of claims 1 to 5 is implemented.

Images