CN-121787871-B - Multi-equipment linkage intelligent scheduling method and system for automobile disassembly production line

CN121787871BCN 121787871 BCN121787871 BCN 121787871BCN-121787871-B

Abstract

The invention relates to the technical field of intelligent manufacturing, and discloses a multi-equipment linkage intelligent scheduling method and system for an automobile disassembly production line. The method comprises the steps of obtaining multi-source operation data to obtain an equipment load distribution matrix and a material flow vector, obtaining a task scheduling priority matrix according to the equipment load distribution matrix and the material flow vector by utilizing a particle swarm algorithm, determining a linkage bottleneck region, outputting a task allocation scheme in a virtual simulation environment by utilizing a deep reinforcement learning model, determining a congestion region according to the task allocation scheme, generating a path correction vector, scheduling available alternative equipment to execute load transfer to obtain a load adjustment parameter, calculating a progress deviation vector according to the load adjustment parameter, and updating the cooperative operation parameter of the multi-equipment. The method can solve the problem of cascade unbalance caused by capability difference of heterogeneous equipment of the dismantling production line.

Inventors

ZHANG LEI
ZHANG LANLAN
WU QIAOZHEN

Assignees

江苏苏北废旧汽车家电拆解再生利用有限公司

Dates

Publication Date: 20260508
Application Date: 20260309

Claims (8)

1. The intelligent scheduling method for the multi-equipment linkage of the automobile disassembly production line is characterized by comprising the following steps of: the method comprises the steps of obtaining multi-source operation data of an automobile disassembly production line, and preprocessing to obtain a device load distribution matrix and a material flow vector; according to the task scheduling priority matrix and a preset processing capacity threshold, determining a blocking node, and determining a linkage bottleneck region and the severity of the linkage bottleneck region according to the blocking node and a preset equipment theoretical flux; If the severity exceeds a preset safety threshold, constructing a virtual simulation environment according to the linkage bottleneck region, performing strategy interaction and search in the virtual simulation environment by utilizing a pre-trained deep reinforcement learning model, and outputting a task allocation scheme; according to the task allocation scheme, calculating the material occupancy rate in a preset workshop space topology, if the material occupancy rate exceeds a preset congestion threshold value, determining a congestion area according to the space distribution of the material occupancy rate, and generating a path correction vector for the congestion area; Scheduling available alternative equipment outside the congestion area to take over tasks to be processed in the congestion area, and executing load transfer to obtain load adjustment parameters; calculating a progress deviation vector by using a preset production progress simulation model according to the load adjustment parameters, and updating the cooperative operation parameters of the multiple devices according to the progress deviation vector; the determining the linkage bottleneck region and the severity of the linkage bottleneck region according to the blocking node and the preset equipment theoretical flux comprises the following steps: acquiring the real-time task accumulation amount of the blocking node, and calculating the difference value between the real-time task accumulation amount and the preset equipment theoretical flux; dividing the difference by the theoretical flux of the device to obtain the quantized severity; Determining the blocking node with the highest severity degree and the adjacent upstream node as the linkage bottleneck region; the method for performing strategy interaction and searching in the virtual simulation environment by utilizing a pre-trained deep reinforcement learning model and outputting a task allocation scheme comprises the following steps: Collecting the length of an equipment queue, the residual processing capacity and the buffer area occupancy rate of the virtual simulation environment, and constructing a state space vector; Outputting an action instruction according to the state space vector by utilizing the strategy network of the deep reinforcement learning model, wherein the action instruction comprises adjusting task priority weight or modifying transmission band beat parameters; Executing the action instruction in the virtual simulation environment, and calculating the yield increase rate in unit time and the key procedure delay reduction rate after execution; and calculating a comprehensive rewarding value according to the yield increasing rate and the delay reducing rate, and updating the strategy network by utilizing the comprehensive rewarding value until the comprehensive rewarding value meets a preset convergence condition, and outputting the task allocation scheme under a corresponding action sequence.
2. The intelligent multi-equipment linkage scheduling method for the automobile disassembly production line according to claim 1, wherein the steps of obtaining multi-source operation data of the automobile disassembly production line, and preprocessing the multi-source operation data to obtain an equipment load distribution matrix and a material flow vector comprise the following steps: Extracting the time stamp of the multi-source operation data, and aligning the data streams with different frequencies to a preset unified time axis by using a linear interpolation algorithm; calculating numerical variance of each aligned data stream in a preset sliding window; Calculating the reciprocal of the numerical variance, and normalizing the reciprocal to obtain a fusion weight; And carrying out weighted summation of index dimensions on the multi-source operation data of each device by utilizing the fusion weight to generate the device load distribution matrix representing the real-time load state of each device.
3. The intelligent scheduling method for multi-equipment linkage of an automobile disassembly production line according to claim 1, wherein the performing iterative optimization by using a particle swarm algorithm according to the equipment load distribution matrix and the material flow vector to obtain a task scheduling priority matrix comprises the following steps: Constructing particle swarms, and defining the position of each particle as a potential task scheduling priority matrix; Defining an adaptability function, wherein the adaptability function is inversely proportional to the average circulation time of the whole process of the production line and inversely proportional to the load variance of each equipment node; and updating the speed and the position of the particles according to the fitness function until the preset iteration times are reached or the preset convergence condition is met, and determining the position of the optimal particles as the task scheduling priority matrix.
4. The method for intelligently scheduling multi-equipment linkage of the automobile disassembly production line according to claim 1, wherein the constructing a virtual simulation environment according to the linkage bottleneck region comprises the following steps: acquiring physical parameters and a current running state of equipment in the linkage bottleneck area; constructing a virtual model consistent with the physical entity geometric parameters according to the equipment physical parameters; Mapping the current running state to the virtual model, and configuring interaction running rules among the virtual models by utilizing the task scheduling priority matrix to generate the virtual simulation environment.
5. The method for intelligent scheduling of multiple equipment linkages in an automobile disassembly line according to claim 1, wherein the generating a path correction vector for the congestion area comprises: dividing the preset workshop space topology into a grid map, and mapping the material occupancy rate into a passing cost weight of a grid; analyzing the task allocation scheme, and determining the starting position and the target processing position of the current material to be processed; Searching a minimum cost path from the starting position to the target processing position in the grid map by using an A-path searching algorithm; And calculating the coordinate difference value between the coordinates of the key node of the minimum cost path and the corresponding node of the original path, and generating the path correction vector.
6. The method for intelligently scheduling multi-equipment linkage of an automobile disassembly production line according to claim 1, wherein the scheduling available substitute equipment located outside the congestion area takes over tasks to be processed in the congestion area, performs load transfer, and obtains load adjustment parameters, and comprises the following steps: Acquiring all candidate equipment sets outside the congestion area; Traversing the candidate device set, and checking the current running state and process attribute of each candidate device; screening out equipment which is idle in the current running state and has the process attribute matched with the task to be processed, and taking the equipment as the available substitute equipment; and generating a scheduling instruction aiming at the available alternative equipment to obtain the load adjustment parameter.
7. The intelligent scheduling method for multi-equipment linkage of an automobile disassembly production line according to claim 1, wherein the calculating a progress deviation vector according to the load adjustment parameter by using a preset production progress simulation model and updating the operation parameters of multi-equipment cooperation according to the progress deviation vector comprises: if the modular length of the progress deviation vector exceeds a preset fault tolerance interval, generating a global beat compensation coefficient for multi-equipment cooperation according to the direction of the progress deviation vector; And correcting the running power or the transmission speed of each device by using the global beat compensation coefficient.
8. Multi-equipment linkage intelligent scheduling system of automobile disassembly production line is characterized by comprising: The data processing module is used for acquiring multi-source operation data of the automobile disassembly production line and preprocessing the multi-source operation data to obtain an equipment load distribution matrix and a material flow vector; The bottleneck identification module is used for carrying out iterative optimization by utilizing a particle swarm algorithm according to the equipment load distribution matrix and the material flow vector to obtain a task scheduling priority matrix, determining a blocking node according to the task scheduling priority matrix and a preset processing capacity threshold value, and determining a linkage bottleneck region and the severity of the linkage bottleneck region according to the blocking node and a preset equipment theoretical flux; the strategy optimization module is used for constructing a virtual simulation environment according to the linkage bottleneck region if the severity exceeds a preset safety threshold value; the path monitoring module is used for calculating the material occupancy rate in the preset workshop space topology according to the task allocation scheme, determining a congestion area according to the space distribution of the material occupancy rate if the material occupancy rate exceeds a preset congestion threshold value, and generating a path correction vector for the congestion area; the load transfer module is used for scheduling available alternative equipment positioned outside the congestion area to take over the task to be processed in the congestion area, and executing load transfer to obtain a load adjustment parameter; The feedback control module is used for calculating a progress deviation vector by using a preset production progress simulation model according to the load adjustment parameters, and updating the cooperative operation parameters of the multiple devices according to the progress deviation vector; the determining the linkage bottleneck region and the severity of the linkage bottleneck region according to the blocking node and the preset equipment theoretical flux comprises the following steps: acquiring the real-time task accumulation amount of the blocking node, and calculating the difference value between the real-time task accumulation amount and the preset equipment theoretical flux; dividing the difference by the theoretical flux of the device to obtain the quantized severity; Determining the blocking node with the highest severity degree and the adjacent upstream node as the linkage bottleneck region; the method for performing strategy interaction and searching in the virtual simulation environment by utilizing a pre-trained deep reinforcement learning model and outputting a task allocation scheme comprises the following steps: Collecting the length of an equipment queue, the residual processing capacity and the buffer area occupancy rate of the virtual simulation environment, and constructing a state space vector; Outputting an action instruction according to the state space vector by utilizing the strategy network of the deep reinforcement learning model, wherein the action instruction comprises adjusting task priority weight or modifying transmission band beat parameters; Executing the action instruction in the virtual simulation environment, and calculating the yield increase rate in unit time and the key procedure delay reduction rate after execution; and calculating a comprehensive rewarding value according to the yield increasing rate and the delay reducing rate, and updating the strategy network by utilizing the comprehensive rewarding value until the comprehensive rewarding value meets a preset convergence condition, and outputting the task allocation scheme under a corresponding action sequence.

Description

Multi-equipment linkage intelligent scheduling method and system for automobile disassembly production line Technical Field The invention relates to the technical field of intelligent manufacturing, in particular to a multi-equipment linkage intelligent scheduling method and system for an automobile disassembly production line. Background Currently, the mainstream automobile disassembly production line generally relies on a traditional industrial control system for job management. In the prior art, a static scheduling mode based on preset beats is adopted in a production line, namely, a central controller sets fixed running speed and start-stop logic for heterogeneous equipment such as a crusher, a magnetic separator, a nonferrous metal separator and the like according to theoretical average working hours, and all the equipment are relatively independent and are simply connected in series only through a physical conveyor belt. However, when facing non-standard input scenes such as large differences in the vehicle conditions of the scrapped vehicles and different rusting degrees of parts, the static scheduling mode is difficult to cope with random fluctuation of time consumption of working procedures. When the processing time of a certain procedure is prolonged or shortened, the fixed transmission beat cannot be dynamically adapted, so that the logistic state and the equipment operation capacity in the production line are disjointed. In the prior art, the technical problems of local congestion and resource idling and coexistence caused by linkage unbalance of multiple equipment loads of an automobile disassembly production line under a dynamic operation scene exist. Disclosure of Invention The invention provides a multi-equipment linkage intelligent scheduling method and system for an automobile disassembly production line, which are used for solving the technical problem that in the prior art, local congestion and resource idling coexist due to multi-equipment load linkage unbalance of the automobile disassembly production line in a dynamic operation scene. In order to solve the technical problems, the invention provides a multi-equipment linkage intelligent scheduling method for an automobile disassembly production line, which comprises the following steps: the method comprises the steps of obtaining multi-source operation data of an automobile disassembly production line, and preprocessing to obtain a device load distribution matrix and a material flow vector; according to the task scheduling priority matrix and a preset processing capacity threshold, determining a blocking node, and determining a linkage bottleneck region and the severity of the linkage bottleneck region according to the blocking node and a preset equipment theoretical flux; If the severity exceeds a preset safety threshold, constructing a virtual simulation environment according to the linkage bottleneck region, performing strategy interaction and search in the virtual simulation environment by utilizing a pre-trained deep reinforcement learning model, and outputting a task allocation scheme; according to the task allocation scheme, calculating the material occupancy rate in a preset workshop space topology, if the material occupancy rate exceeds a preset congestion threshold value, determining a congestion area according to the space distribution of the material occupancy rate, and generating a path correction vector for the congestion area; Scheduling available alternative equipment outside the congestion area to take over tasks to be processed in the congestion area, and executing load transfer to obtain load adjustment parameters; And calculating a progress deviation vector by using a preset production progress simulation model according to the load adjustment parameters, and updating the cooperative operation parameters of the multiple devices according to the progress deviation vector. In a second aspect, the invention provides a multi-device linkage intelligent scheduling system for an automobile disassembly production line, comprising: The data processing module is used for acquiring multi-source operation data of the automobile disassembly production line and preprocessing the multi-source operation data to obtain an equipment load distribution matrix and a material flow vector; The bottleneck identification module is used for carrying out iterative optimization by utilizing a particle swarm algorithm according to the equipment load distribution matrix and the material flow vector to obtain a task scheduling priority matrix, determining a blocking node according to the task scheduling priority matrix and a preset processing capacity threshold value, and determining a linkage bottleneck region and the severity of the linkage bottleneck region according to the blocking node and a preset equipment theoretical flux; the strategy optimization module is used for constructing a virtual simulation environment according to the linkage b