Digital-analog double-drive intelligent scheduling method and device for cross-organization machining workshop
Through the digital-to-analog dual-drive intelligent scheduling method, combined with genetic evolution and data-driven strategies, cross-organizational scheduling of the machine-added workshop is solved, and the problem of inefficient scheduling in the existing technology is achieved and more efficient resource allocation and production efficiency is achieved.
Patent Information
- Application Number
- CN202410984633.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-07-22
AI Technical Summary
The existing cross-organization workshop scheduling methods are difficult to record historical search experience while achieving fast and efficient search, and the solution to the machining workshop scheduling problems with complex resource grouping is low.
Using digital-to-analog dual-drive intelligent scheduling method, a cross-organization machine-to-shop scheduling model is established by modeling the production scenario of the machine-to-shop production scenario, and combining genetic evolution search and data-driven neighborhood search strategies, populations are generated and search knowledge is recorded, and the search knowledge is used. Neural network optimization scheduling scheme.
It improves the solution efficiency of scheduling problems in complex machining workshops for resource grouping, reduces the manufacturing cost of aviation products, and solves the problem that it is difficult for the existing technology to quickly and efficiently search and record historical search experience.
Smart Images

Figure CN118941017A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of cross-organizational workshop scheduling, and in particular to a digital-analog dual-drive intelligent scheduling method and device for a cross-organizational machining workshop. Background Art
[0002] The machining workshop is an important production environment in the process route of aviation products. The machining workshop is responsible for cutting and processing the hot-pressed composite materials into the required shape and size. Modern manufacturing companies adopt a multi-park distributed manufacturing model. After the order is placed at the headquarters, the command center directs the transportation fleet to send the raw materials from the warehouse to different parks. Different parks have different processing capabilities, but they can be processed in parallel. In order to increase production capacity, companies choose to distribute raw materials to multiple parks for processing, which increases the complexity of the problem.
[0003] The existing technologies that consider cross-organizational workshop scheduling problems still have obvious deficiencies at the methodological level.
[0004] At the method level, the current solutions to distributed workshop scheduling problems can be divided into data-driven and knowledge-driven. Among them, knowledge-driven algorithms generally use intelligent optimization algorithms combined with knowledge-driven local search operations. This method can achieve fast global search, but it cannot record the search knowledge of each generation. The algorithm directly eliminates the old population after the population update iteration is completed, but the eliminated population contains historical search experience, which can be effectively used to provide knowledge for subsequent searches of the algorithm; data-driven algorithms generally use reinforcement learning algorithms. This type of algorithm uses neural networks to store the search strategy of each generation, but this type of algorithm has a slow search speed, random strategy search direction, and lacks guidance.
[0005] In summary, the existing cross-organizational workshop scheduling methods are difficult to achieve fast and efficient search while recording historical search experience. The efficiency of solving the scheduling problem of machining workshops with complex resource grouping is low and needs to be solved urgently. Summary of the invention
[0006] The present application provides a digital-analog dual-drive intelligent scheduling method and device for a cross-organizational machine shop to solve the problems of the existing cross-organizational workshop scheduling methods, such as difficulty in achieving fast and efficient search while recording historical search experience, and low efficiency in solving scheduling problems of machine shops with complex resource grouping.
[0007] The first aspect of the present application provides a digital-analog dual-drive intelligent scheduling method for a cross-organizational machine shop, comprising the following steps: modeling an actual machine shop production scenario, establishing a cross-organizational machine shop scheduling model corresponding to the machine shop production scenario, and vector encoding the cross-organizational machine shop scheduling model to obtain a factory vector, a process vector, and a machine vector corresponding to the cross-organizational machine shop scheduling model; generating a first parent population and a second parent population corresponding to the cross-organizational machine shop scheduling model according to a preset collaborative initialization strategy, and performing a genetic evolution search operation on the first parent population to obtain a first child population of the first parent population, and based on the second parent population and a preset data-driven neighborhood search strategy, Acquire a second child population and record the first search knowledge corresponding to the second child population; generate a new first parent population based on the first parent population and the second child population, and obtain a new second parent population based on the second parent population and the first child population, and perform a multi-operator collaborative neighborhood search operation on the new first parent population to obtain the second search knowledge corresponding to the new first parent population; horizontally splice the factory vector, the process vector and the machine vector according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and input the one-dimensional tensor into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets preset requirements corresponding to the production scenario of the machining workshop.
[0008] Optionally, in one embodiment of the present application, the preset collaborative initialization strategy includes: using the cross-organizational machine shop scheduling model to select a parallel machine with the shortest processing time for each workpiece; or, accumulating the processing time of each workpiece in different factories to obtain a cumulative value corresponding to each workpiece, and arranging the cumulative values of all workpieces in descending order to generate a cumulative value sequence, so as to allocate each workpiece to a corresponding factory according to the cumulative value sequence; or, for each workpiece, selecting a parallel machine with the shortest completion time among all factories; or, according to the delivery date of each process in the cross-organizational machine shop scheduling model, arranging each process in ascending order to obtain a process sequence, and allocating each workpiece to a corresponding factory based on the process sequence and a preset factory code.
[0009] Optionally, in one embodiment of the present application, generating the first parent population and the second parent population corresponding to the cross-organizational machine shop scheduling model according to a preset collaborative initialization strategy includes: calculating the total delay and the maximum completion time corresponding to the machine shop production scenario according to the factory vector, the process vector and the machine vector; randomly initializing multiple target sub-populations, and adjusting the factory vector, the process vector and the machine vector corresponding to each target sub-population in the multiple target sub-populations through any one of the preset multiple initialization strategies; merging the multiple target sub-populations according to the adjusted factory vector, the process vector and the machine vector to obtain the first parent population and the second parent population.
[0010] Optionally, in one embodiment of the present application, the performing a genetic evolutionary search operation on the first parent population to obtain a first child population of the first parent population includes: based on a two-person tournament algorithm, obtaining a mating pool corresponding to the first parent population, and randomly selecting a parent from the mating pool through each individual in the first parent population; exchanging the process processing sequence of the cross-organizational machine shop scheduling model with the parents according to a process-based crossover strategy, and using a general crossover strategy to exchange the machine selection information and factory information of the cross-organizational machine shop scheduling model to generate a first child population of the first parent population.
[0011] Optionally, in one embodiment of the present application, the preset data-driven neighborhood search strategy includes: searching for the key workpiece with the largest delay in the cross-organizational machine shop scheduling model, and searching forward for the first target workpiece that meets the preset delivery period requirements, exchanging the positions of the key workpiece and the first target workpiece, or inserting the key workpiece forward to the front position of the first target workpiece; or, randomly selecting any workpiece in the cross-organizational machine shop scheduling model, and searching for the second target workpiece that meets the preset delivery period requirements, and exchanging the positions of the second target workpiece and any workpiece; or, searching for the key factory of the cross-organizational machine shop scheduling model and the key path corresponding to the key factory according to the maximum completion time, and randomly selecting a key process at the same time, and scanning the completion time of all machines to obtain the target machine that meets the preset completion time requirements, so as to insert the key process into the target machine.
[0012] Optionally, in one embodiment of the present application, the one-dimensional tensor is input into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets preset requirements corresponding to the production scenario of the machining workshop, including: extracting the current state of the one-dimensional tensor, and inputting the current state into the evaluation network of the neighborhood action recommendation neural network model to obtain the Q value of the current state, and determining the Q value of the current strategy according to the selected neighborhood action; extracting the next state of the one-dimensional tensor, and inputting the next state into the target network of the neighborhood action recommendation neural network model to obtain the Q value of all actions in the next state; based on the Q value of all actions in the next state, The Q value of the action is obtained, the maximum Q value of all actions in the next state is obtained, and the target value is calculated according to the maximum Q value, the preset discount factor and the reward information of the one-dimensional tensor; the root mean square error of the neighborhood action recommendation neural network model is calculated according to the Q value of the current action and the target value, and the evaluation network parameters are optimized by a preset back propagation algorithm, and when the target network parameters meet the preset update requirements, the target network parameters are updated by using the optimized evaluation network parameters; the root mean square error is adjusted by the updated target network parameters, and when the root mean square error meets the preset error condition, a scheduling plan that meets the preset requirements corresponding to the production scenario of the machining workshop is obtained.
[0013] The second aspect of the present application provides a digital-analog dual-drive intelligent scheduling device for a cross-organizational machine shop, including: a modeling module, used to model the actual machine shop production scene, establish a cross-organizational machine shop scheduling model corresponding to the machine shop production scene, and vector encode the cross-organizational machine shop scheduling model to obtain the factory vector, process vector and machine vector corresponding to the cross-organizational machine shop scheduling model; a data-driven module, used to generate a first parent population and a second parent population corresponding to the cross-organizational machine shop scheduling model according to a preset collaborative initialization strategy, and perform a genetic evolution search operation on the first parent population to obtain a first child population of the first parent population, and at the same time obtain a first child population based on the second parent population and a preset data-driven neighborhood search strategy. a second child population, and records the first search knowledge corresponding to the second child population; a knowledge-driven module, for generating a new first parent population based on the first parent population and the second child population, and obtaining a new second parent population according to the second parent population and the first child population, and performing a multi-operator collaborative neighborhood search operation on the new first parent population to obtain the second search knowledge corresponding to the new first parent population; a splicing module, for horizontally splicing the factory vector, the process vector and the machine vector according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and inputting the one-dimensional tensor into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets preset requirements corresponding to the production scenario of the machining workshop.
[0014] Optionally, in one embodiment of the present application, the data-driven module includes: a first initialization strategy unit, used to select a parallel machine with the shortest processing time for each workpiece in the cross-organizational machine shop scheduling model; or, a second initialization strategy unit, used to accumulate the processing time of each workpiece in different factories to obtain the cumulative value corresponding to each workpiece, and arrange the cumulative values of all workpieces in descending order to generate a cumulative value sequence, so as to allocate each workpiece to the corresponding factory according to the cumulative value sequence; or, a third initialization strategy unit, used to select a parallel machine with the shortest completion time in all factories for each workpiece; or, a fourth initialization strategy unit, used to arrange each process in ascending order according to the delivery date of each process in the cross-organizational machine shop scheduling model to obtain a process sequence, and allocate each workpiece to the corresponding factory based on the process sequence and a preset factory code.
[0015] Optionally, in one embodiment of the present application, the data-driven module also includes: a calculation unit, used to calculate the total delay and maximum completion time corresponding to the machining workshop production scenario based on the factory vector, the process vector and the machine vector; an adjustment unit, used to randomly initialize multiple target sub-populations, and adjust the factory vector, the process vector and the machine vector corresponding to each target sub-population in the multiple target sub-populations through any one of multiple preset initialization strategies; a merging unit, used to merge the multiple target sub-populations according to the adjusted factory vector, the process vector and the machine vector to obtain the first parent population and the second parent population.
[0016] Optionally, in one embodiment of the present application, the data-driven module also includes: a selection unit, used to obtain the mating pool corresponding to the first parent population based on a two-player tournament algorithm, and randomly select parents from the mating pool through each individual in the first parent population; an exchange unit, used to exchange the process processing sequence of the cross-organizational machine shop scheduling model with the parent according to a process-based crossover strategy, and use a general crossover strategy to exchange the machine selection information and factory information of the cross-organizational machine shop scheduling model to generate a first child population of the first parent population.
[0017] Optionally, in one embodiment of the present application, the data-driven module also includes: a first neighborhood search unit, used to find the key workpiece with the largest delay in the cross-organizational machine shop scheduling model, and forward search for the first target workpiece that meets the preset delivery period requirements, exchange the positions of the key workpiece and the first target workpiece, or insert the key workpiece forward to the front position of the first target workpiece; or, a second neighborhood search unit, used to randomly select any workpiece in the cross-organizational machine shop scheduling model, and search for the second target workpiece that meets the preset delivery period requirements, and exchange the positions of the second target workpiece and any of the workpieces; or, a third neighborhood search unit, used to find the key factory and the key path corresponding to the key factory in the cross-organizational machine shop scheduling model according to the maximum completion time, and randomly select a key process at the same time, and scan the completion time of all machines to obtain the target machine that meets the preset completion time requirements, so as to insert the key process into the target machine.
[0018] Optionally, in one embodiment of the present application, the splicing module includes: a first extraction unit, used to extract the current state of the one-dimensional tensor, and input the current state into the evaluation network of the neighborhood action recommendation neural network model to obtain the Q value of the current state, and determine the Q value of the current strategy according to the selected neighborhood action; a second extraction unit, used to extract the next state of the one-dimensional tensor, and input the next state into the target network of the neighborhood action recommendation neural network model to obtain the Q values of all actions in the next state; an acquisition unit, used to obtain the maximum Q value of all actions in the next state based on the Q values of all actions in the next state Q value, and calculate the target value according to the maximum Q value, the preset discount factor and the reward information of the one-dimensional tensor; an optimization unit, used to calculate the root mean square error of the neighborhood action recommendation neural network model according to the Q value of the current action and the target value, and optimize the evaluation network parameters through a preset back propagation algorithm, and at the same time, when the target network parameters meet the preset update requirements, the target network parameters are updated using the optimized evaluation network parameters; an adjustment unit, used to adjust the root mean square error through the updated target network parameters, and when the root mean square error meets the preset error condition, obtain a scheduling plan that meets the preset requirements corresponding to the production scenario of the machining workshop.
[0019] The third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the digital-analog dual-drive intelligent scheduling method for cross-organizational machine shops as described in the above embodiment.
[0020] The fourth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above-mentioned digital-analog dual-drive intelligent scheduling method for cross-organizational machine shops.
[0021] Therefore, the embodiments of the present application have the following beneficial effects:
[0022] The embodiments of the present application can establish a cross-organizational machine shop scheduling model corresponding to the machine shop production scenario by modeling the actual machine shop production scenario, and vector encode the cross-organizational machine shop scheduling model to obtain the factory vector, process vector and machine vector corresponding to the cross-organizational machine shop scheduling model; generate the first parent population and the second parent population corresponding to the cross-organizational machine shop scheduling model according to the preset collaborative initialization strategy, and perform a genetic evolution search operation on the first parent population to obtain the first child population of the first parent population, and at the same time obtain the second child population based on the second parent population and the preset data-driven neighborhood search strategy , and record the first search knowledge corresponding to the second child population; based on the first parent population and the second child population, generate a new first parent population, and at the same time, according to the second parent population and the first child population, obtain a new second parent population, and perform a multi-operator collaborative neighborhood search operation on the new first parent population to obtain the second search knowledge corresponding to the new first parent population; according to the first search knowledge and the second search knowledge, horizontally splice the factory vector, process vector and machine vector to obtain a one-dimensional tensor, and input the one-dimensional tensor into the pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets the preset requirements corresponding to the production scene of the machining workshop. This application models the production scene of the actual machining workshop as a cross-organizational machining workshop scheduling problem by considering various resource configurations and constraints, thereby improving the manufacturing capacity of aviation products in the machining workshop from the perspective of resource optimization, improving the efficiency of solving the scheduling problem of the machining workshop with complex resource grouping, and reducing the manufacturing cost of aviation products. As a result, it solves the problems that the existing cross-organizational workshop scheduling method is difficult to record historical search experience while achieving fast and efficient search, and the efficiency of solving the scheduling problem of the machining workshop with complex resource grouping is low.
[0023] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:
[0025] Figure 1 It is a flow chart of a digital-analog dual-drive intelligent scheduling method for a cross-organizational machining workshop provided according to an embodiment of the present application;
[0026] Figure 2 A Gantt chart of a cross-organizational machine shop scheduling model provided for one embodiment of the present application;
[0027] Figure 3 A schematic diagram of an N1-N3 neighborhood structure provided for an embodiment of the present application;
[0028] Figure 4 A schematic diagram of an N4 neighborhood structure provided for an embodiment of the present application;
[0029] Figure 5 A schematic diagram of a feedforward neural network provided for one embodiment of the present application;
[0030] Figure 6 A schematic diagram of a deep Q network algorithm training process provided for an embodiment of the present application;
[0031] Figure 7 A schematic diagram of the execution logic of a digital-analog dual-drive intelligent scheduling method for a cross-organizational machining workshop provided by one embodiment of the present application;
[0032] Figure 8 A schematic diagram of the logical architecture of a digital-analog dual-drive intelligent scheduling method for a cross-organizational machining workshop provided for one embodiment of the present application;
[0033] Fig. 9 This is an example diagram of a digital-analog dual-drive intelligent scheduling device for a cross-organizational machining workshop according to an embodiment of the present application;
[0034] Fig.10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.
[0035] Among them, 10-digital-analog dual-drive intelligent scheduling device for cross-organizational machine processing workshops; 100-modeling module, 200-data driving module, 300-knowledge driving module, 400-splicing module; 1001-memory, 1002-processor, 1003-communication interface. DETAILED DESCRIPTION
[0036] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.
[0037] The following describes the digital-analog dual-drive intelligent scheduling method and device for a cross-organizational machine shop according to an embodiment of the present application with reference to the accompanying drawings. In response to the problems mentioned in the above background technology, the present application provides a digital-analog dual-drive intelligent scheduling method for a cross-organizational machine shop. In this method, by modeling the actual production scene of the machine shop, a cross-organizational machine shop scheduling model corresponding to the production scene of the machine shop is established, and the cross-organizational machine shop scheduling model is vector-encoded to obtain the factory vector, process vector and machine vector corresponding to the cross-organizational machine shop scheduling model; the first parent population and the second parent population corresponding to the cross-organizational machine shop scheduling model are generated according to the preset collaborative initialization strategy, and a genetic evolution search operation is performed on the first parent population to obtain the first child population of the first parent population, and at the same time, based on the second parent population and the preset A data-driven neighborhood search strategy obtains the second child population and records the first search knowledge corresponding to the second child population; based on the first parent population and the second child population, a new first parent population is generated, and at the same time, a new second parent population is obtained based on the second parent population and the first child population, and a multi-operator collaborative neighborhood search operation is performed on the new first parent population to obtain the second search knowledge corresponding to the new first parent population; according to the first search knowledge and the second search knowledge, the factory vector, process vector and machine vector are horizontally spliced to obtain a one-dimensional tensor, and the one-dimensional tensor is input into the pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets the preset requirements corresponding to the production scenario of the machining workshop. This application models the production scenario of the actual machining workshop as a cross-organizational machining workshop scheduling problem by considering various resource configurations and constraints, thereby improving the manufacturing capacity of aviation products in the machining workshop from the perspective of resource optimization, improving the efficiency of solving the scheduling problem of the machining workshop with complex resource grouping, and reducing the manufacturing cost of aviation products. This solves the problems of existing cross-organizational workshop scheduling methods, such as difficulty in achieving fast and efficient search while recording historical search experience, and low efficiency in solving scheduling problems for machining workshops with complex resource grouping.
[0038] Specifically, Figure 1 A flowchart of a digital-analog dual-drive intelligent scheduling method for a cross-organizational machine shop provided in an embodiment of the present application.
[0039] like Figure 1 As shown, the digital-analog dual-drive intelligent scheduling method for the cross-organizational machining workshop includes the following steps:
[0040] In step S101, the actual machine shop production scenario is modeled, a cross-organizational machine shop scheduling model corresponding to the machine shop production scenario is established, and the cross-organizational machine shop scheduling model is vector encoded to obtain the factory vector, process vector and machine vector corresponding to the cross-organizational machine shop scheduling model.
[0041] Those skilled in the art should understand that in order to improve the manufacturing capacity of aviation products in machine shops from the perspective of resource optimization, the embodiments of the present application model the actual production scenario of a machine shop, and take into account various resource configurations and constraints, and ultimately model the actual production scenario of a machine shop as a cross-organizational machine shop scheduling model.
[0042] It should be noted that if Figure 2 As shown, the cross-organizational machining workshop scheduling model in the embodiment of the present application can be described as follows: N workpieces need to be processed, each workpiece has only one process, namely the machining process; there are a total of M parallel machines in the workshop, a total of F parks, and each group is responsible for using M f parallel machines, each parallel machine requires one worker to start processing; the optimization goal of the cross-organizational machine shop scheduling model is to minimize the total delivery tardiness (Total tardiness, TT) and the maximum completion time C simultaneously. max , its calculation expression is:
[0043]
[0044]
[0045] Among them, C i represents the completion time of order i; D i Indicates the delivery period.
[0046] In addition, the embodiments of the present application may also adopt a three-layer vector encoding strategy to vector encode the cross-organizational machine shop scheduling model to cover all solutions in the solution space of the model. Specifically, the embodiments of the present application may first use a permutation of 1 to N to represent the order of workpieces, for example, [5,3,1,2,4] represents the processing order of workpieces 1-5; secondly, a factory vector of length N is used to represent the park selection information of the workpiece, and each dimension is labeled with the workpiece grouping information, for example, [1,2,1,2,2] represents the grouping information of workpieces 1-5; finally, a machine vector of length N is used to represent the machine selection of the workpiece, for example, [2,1,2,2,1] represents the machine selection information of workpieces 1-5, where workpiece 1 selects machine 2.
[0047] Therefore, the embodiments of the present application model the actual machine shop production scenario as a cross-organizational machine shop scheduling model and vector encode the cross-organizational machine shop scheduling model, thereby providing a reliable data basis for the execution of subsequent multi-strategy collaborative initialization and other operations.
[0048] In step S102, the first parent population and the second parent population corresponding to the cross-organizational machine shop scheduling model are generated according to the preset collaborative initialization strategy, and a genetic evolutionary search operation is performed on the first parent population to obtain the first child population of the first parent population. At the same time, based on the second parent population and the preset data-driven neighborhood search strategy, the second child population is obtained, and the first search knowledge corresponding to the second child population is recorded.
[0049] Furthermore, the embodiments of the present application also need to generate the first parent population and the second parent population corresponding to the cross-organizational machine shop scheduling model through a collaborative initialization strategy; in addition, the embodiments of the present application also set up search strategies of genetic evolutionary search and knowledge-driven local search for rapid convergence in the solution space.
[0050] During the actual execution process, the embodiments of the present application can perform a genetic evolution search operation on the first parent population to obtain the first child population of the first parent population. At the same time, the embodiments of the present application also need to perform a data-driven neighborhood search process on the second parent population to generate a second child population and store the first search knowledge corresponding to the second child population.
[0051] Optionally, in one embodiment of the present application, a collaborative initialization strategy is preset, including: selecting a parallel machine with the shortest processing time for each workpiece in a cross-organizational machine shop scheduling model; or, accumulating the processing time of each workpiece in different factories to obtain a cumulative value corresponding to each workpiece, and arranging the cumulative values of all workpieces in descending order to generate a cumulative value sequence, so as to allocate each workpiece to a corresponding factory according to the cumulative value sequence; or, selecting a parallel machine with the shortest completion time among all factories for each workpiece; or, arranging each process in ascending order according to the delivery date of each process in the cross-organizational machine shop scheduling model to obtain a process sequence, and allocating each workpiece to a corresponding factory based on the process sequence and a preset factory code.
[0052] It should be noted that the collaborative initialization strategy in the embodiment of the present application includes the following four initialization strategy forms:
[0053] 1. Select the parallel machine with the shortest processing time for all workpieces, and reduce the completion time by reducing the processing time;
[0054] 2. Accumulate the processing time of all workpieces in different factories, sort them in descending order according to the accumulated value, and distribute them to different factories for processing in this order to ensure that the machine with the longest processing time is processed first;
[0055] 3. Each job selects the parallel machine with the minimum completion time among all factories;
[0056] 4. Arrange all processes in ascending order according to the delivery date, and then assign the workpieces to different factories according to the factory code FA, so that the workpieces with earlier delivery dates can be processed first, ultimately reducing the total delay.
[0057] Therefore, the embodiments of the present application provide a solid theoretical basis for high-quality initial populations by determining the above-mentioned collaborative initialization strategy.
[0058] Optionally, in one embodiment of the present application, a first parent population and a second parent population corresponding to a cross-organizational machine shop scheduling model are generated according to a preset collaborative initialization strategy, including: calculating a total delay and a maximum completion time corresponding to a machine shop production scenario according to a factory vector, a process vector, and a machine vector; randomly initializing multiple target sub-populations, and adjusting the factory vector, the process vector, and the machine vector corresponding to each target sub-population in the multiple target sub-populations through any of a plurality of preset initialization strategies; merging the multiple target sub-populations according to the adjusted factory vector, the process vector, and the machine vector to obtain a first parent population and a second parent population.
[0059] After obtaining the encoding vector, the embodiment of the present application can divide the workpieces into different shifts according to the factory vector, and extract the processes in order according to the process vector; secondly, obtain the machine tool information of the workpiece according to the machine vector, arrange the workpiece to be processed on the corresponding machine and calculate the start time and completion time; finally, after all workpieces are processed, calculate the delay of each workpiece, accumulate all delays to obtain the total delay TT, and at the same time obtain the maximum completion time C max .
[0060] Afterwards, the embodiments of the present application can randomly initialize multiple target sub-populations, such as randomly initializing 5 sub-populations, and their sizes are ps / 5 respectively; secondly, for each sub-population, an initialization strategy is used to adjust different vector encodings (i.e., factory vector, process vector and machine vector); finally, multiple target sub-populations are merged into one population, whose size is ps, thereby obtaining the first parent population and the second parent population through this initialization method.
[0061] Therefore, the embodiment of the present application effectively enhances the convergence of the initial population by initializing the sub-population using the above-mentioned collaborative initialization strategy, thereby integrating into a high-quality initial population.
[0062] Optionally, in one embodiment of the present application, a genetic evolutionary search operation is performed on the first parent population to obtain a first child population of the first parent population, including: based on a two-person tournament algorithm, obtaining a mating pool corresponding to the first parent population, and randomly selecting a parent from the mating pool through each individual in the first parent population; exchanging the process processing sequence of the cross-organizational machine shop scheduling model with the parents according to a process-based crossover strategy, and using a general crossover strategy to exchange the machine selection information and factory information of the cross-organizational machine shop scheduling model to generate a first child population of the first parent population.
[0063] In the specific implementation process, the embodiment of the present application can select a mating pool from the first parent population based on a genetic evolutionary search strategy using a two-person tournament algorithm; then use individuals in the first parent population to randomly select parents from the mating pool, and exchange the process processing order through a process-based crossover (Precedence Operation Crossover, POX), while using a universal crossover (Universal Crossover, UX) to exchange machine selection information and factory information, thereby generating a first child population of the first parent population.
[0064] Optionally, in one embodiment of the present application, a data-driven neighborhood search strategy is preset, including: finding the critical workpiece with the largest delay in the cross-organizational machine shop scheduling model, and forward searching for the first target workpiece that meets the preset delivery period requirements, exchanging the positions of the critical workpiece and the first target workpiece, or inserting the critical workpiece forward to the front position of the first target workpiece; or, randomly selecting any workpiece in the cross-organizational machine shop scheduling model, and searching for the second target workpiece that meets the preset delivery period requirements, and exchanging the positions of the second target workpiece and any workpiece; or, searching for the critical factory and the critical path corresponding to the critical factory in the cross-organizational machine shop scheduling model according to the maximum completion time, and randomly selecting a critical process at the same time, and scanning the completion time of all machines to obtain the target machine that meets the preset completion time requirements, so as to insert the critical process into the target machine.
[0065] It should be noted that the data-driven neighborhood search strategy of the embodiment of the present application provides four neighborhood structures N1-N4. Specifically, the above four neighborhood structures are as follows:
[0066] N1: Find the key workpiece with the largest delay, search forward for the workpiece with a later delivery date (i.e., the first target workpiece), and exchange the positions of the key workpiece and the selected workpiece (i.e., the first target workpiece);
[0067] N2: Find the key workpiece with the largest delay, find the workpiece with a later delivery date, and insert the key workpiece in front of the selected workpiece (i.e., the first target workpiece);
[0068] N3: If Figure 3 As shown, the embodiment of the present application randomly selects a workpiece, searches for a workpiece with a later delivery date (i.e., the second target workpiece), and exchanges the positions of the two workpieces;
[0069] N4: If Figure 4 As shown, the critical factory is found according to the maximum completion time, the critical path is found at the same time, a key process is randomly selected, the completion time of all machines is scanned, and the machine that does not increase the maximum completion time is selected to be inserted, and finally the maximum completion time is reduced.
[0070] Therefore, the embodiments of the present application design four knowledge-driven neighborhood structures through a knowledge-driven local search strategy, thereby effectively improving the resource optimization efficiency of the Gantt chart and accelerating the convergence of the evolutionary computing population.
[0071] In step S103, a new first parent population is generated based on the first parent population and the second child population, and a new second parent population is obtained based on the second parent population and the first child population. A multi-operator collaborative neighborhood search operation is performed on the new first parent population to obtain second search knowledge corresponding to the new first parent population.
[0072] In step S104, the factory vector, process vector and machine vector are horizontally spliced according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and the one-dimensional tensor is input into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling plan that meets the preset requirements corresponding to the production scenario of the machining workshop.
[0073] Furthermore, the embodiment of the present application can horizontally splice the three encoded vectors (i.e., the factory vector, the process vector, and the machine vector) into a one-dimensional tensor based on the Markov decision strategy, and input it into Figure 5 In the feedforward neural network shown, the scheduling scheme features are extracted, and the selection probabilities of different neighborhood search operators are output based on the scheduling scheme features.
[0074] Among them, the specific design of Markov decision is as follows:
[0075] 1. State: In the embodiments of the present application, each solution represents a state, which can be recorded by the parameters of the feedforward neural network;
[0076] 2. Action: Using knowledge-driven neighborhood structure as the agent to select actions;
[0077] 3. Reward: When action A is selected t Can optimize the current state S t Find a better solution t+1 , then the reward Rt=10, otherwise Rt=0;
[0078] 4. Transfer: The current solution state is S t , execute neighborhood action A t After that, the scheduling information such as processing sequence, machine tool selection, and factory selection has changed, and the solution state has also changed to S t+1
[0079] Therefore, the embodiments of the present application use data-driven multi-operation collaborative search, adopt neural networks to record search information, and train network parameters through reinforcement learning to provide decisions for the evolutionary computing population, thereby solving the problem that the evolutionary algorithm of the traditional method cannot remember historical search experience.
[0080] Optionally, in one embodiment of the present application, a one-dimensional tensor is input into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets preset requirements corresponding to a machine shop production scenario, including: extracting the current state of the one-dimensional tensor, and inputting the current state into the evaluation network of the neighborhood action recommendation neural network model to obtain the Q value of the current state, and determining the Q value of the current strategy based on the selected neighborhood action; extracting the next state of the one-dimensional tensor, and inputting the next state into the target network of the neighborhood action recommendation neural network model to obtain the Q value of all actions in the next state; based on all actions in the next state, The Q value of the next state is obtained, and the maximum Q value of all actions in the next state is obtained, and the target value is calculated according to the maximum Q value, the preset discount factor and the reward information of the one-dimensional tensor; the root mean square error of the neighborhood action recommendation neural network model is calculated according to the Q value and the target value of the current action, and the evaluation network parameters are optimized through the preset back propagation algorithm. At the same time, when the target network parameters meet the preset update requirements, the target network parameters are updated with the optimized evaluation network parameters; the root mean square error is adjusted through the updated target network parameters, and when the root mean square error meets the preset error condition, the scheduling plan that meets the preset requirements corresponding to the production scenario of the machining workshop is obtained.
[0081] As a feasible method, the neighborhood action recommendation neural network in the embodiment of the present application may adopt a deep Q network algorithm, which includes two networks: an evaluation network Q_val and a target network Q_tar. The evaluation network is used to evaluate the Q value of the current state, and the target network is used to fit Q_val and update the parameters through back propagation. The specific training process is as follows:
[0082] 1. Take out data of size bs from the experience pool. The data format is (S t ,A t ,R t ,S t+1 );
[0083] 2. Extraction status S t , input the evaluation network Q_val to get the Q value of the current state, such as Figure 6 As shown, select the Q value Q_val(S t ,A t );
[0084] 3. Extract the next state S t+1 Input the target network Q_tar to get the Q value of all actions in the next state, Q_tar(S t+1 ,:);
[0085] 4. Select the maximum Q value of the next state multiplied by the discount factor γ, plus the reward R t , and get the target value q t .
[0086] 5. Calculate the root mean square error MSE(Q_val(S t ,A t ), q t ), and use the back propagation algorithm to update the parameters of Q_val to minimize the MSE;
[0087] 6. If Q_tar is not updated by iter_max generation, copy the value of Q_val to Q_tar.
[0088] Therefore, the embodiments of the present application establish a cross-organizational machine shop scheduling model based on the actual production characteristics of cross-organizational machine shops for complex aviation products, and form a framework for collaborative solution of genetic evolutionary computing and reinforcement learning through a scheduling algorithm driven by both data and knowledge, thereby achieving rapid convergence and efficient solution to the scheduling problem of machine shops with complex resource grouping.
[0089] The following is an explanation of the execution logic of the digital-analog dual-drive intelligent scheduling method for cross-organizational machining workshops of the present application through a specific embodiment and in combination with the accompanying drawings.
[0090] Figure 7 Schematic diagram of the execution logic of the cross-organizational machining workshop digital-analog dual-drive intelligent scheduling method of this application. Figure 7 As shown, the execution process of the digital-analog dual-drive intelligent scheduling method for cross-organizational machining workshops of the present application is as follows:
[0091] S701: Initialize evolutionary computing and reinforcement learning populations P1 and P2, and neural network parameters;
[0092] S702: Decode populations P1 and P2, and calculate the objective function value;
[0093] S703: Genetic evolution searches population P1 to obtain offspring C1;
[0094] S704: Use a neural network to perform a neighborhood search on P2 to obtain a child C2, and store the search knowledge (P2, A t,R t ,C2);
[0095] S705: Combine P1 and C2, such as Figure 8 As shown, a new population P1' is selected using fast non-dominated sorting;
[0096] S706: Combine P2 and C1, and use fast non-dominated sorting to select a new population P2';
[0097] S707: Perform multi-operator collaborative neighborhood search on population P1' and store search knowledge (P1', A t ,R t ,C1');
[0098] S708: Determine whether the end condition is met, if yes, end, otherwise go to S703.
[0099] In addition, the present application can illustrate and analyze the performance of the digital-analog dual-drive intelligent scheduling method for cross-organizational machining workshops of the present application through a specific embodiment.
[0100] Specifically, in order to verify the performance of the present application in the scheduling of complex welding workshops with worker constraints, the present application can randomly generate 20 test problems based on actual production data for testing. Number of workpieces N = {20, 40, 60, 80, 100}, number of parallel machines M f ={4,6}, number of factories F ={2,3}, and processing time is randomly selected from U to [5,95].
[0101] As a feasible method, the specific embodiment of the present application can use python for related programming operations, the simulation environment is Intel (R) core (TM) i9-13900k CPU @ 3.4GHz / 16GB RAM, GPU NVIDIA407012GB, and the operating system is windows 10; the termination criterion of the algorithm DRLMCCA corresponding to the digital-analog dual-drive intelligent scheduling method of the cross-organizational machine shop of the present application is that the algorithm stops when the maximum number of evaluations MaxNFEs of the algorithm reaches 400*N; in the experiment, hypervolume (Hypervolume, HV) and generation distance (Generation Distance, GD) can be used to represent the algorithm to verify the comprehensive performance and convergence of the present application.
[0102] In order to verify the effectiveness of this application, the following four variant algorithms were designed in the experiment:
[0103] 1. MA-L means evolutionary computation without local search;
[0104] 2. MA-I stands for evolutionary computation deletion multi-strategy collaborative initialization;
[0105] 3. MA means using only evolutionary computing population for solution;
[0106] 4. DRL means using only reinforcement learning population to solve the problem.
[0107] The specific embodiments of the present application can compare the average HV value of DRLMCCA with that of the variant algorithm, as shown in Table 1, and compare the average GD value of DRLMCCA with that of the variant algorithm, as shown in Table 2.
[0108] Table 1
[0109]
[0110]
[0111] Table 2
[0112]
[0113]
[0114] It can be seen from Tables 1 and 2 that DRLMCCA achieved the best HV and GD indicators, indicating that after deleting the improvement strategy and various functional modules, the algorithm performance has significantly decreased, which effectively verifies the effectiveness of each algorithm unit in this application.
[0115] Afterwards, the present application may rank the Frideman rank sum test results of the DRLMCCA and variant algorithms, as shown in Table 3.
[0116] Table 3
[0117]
[0118] As can be seen from Table 3, the highest HV and GD index rankings of DRLMCCA are less than 0.05 compared with other algorithms. After deleting the improvement strategy, the algorithm performance decreases significantly, which verifies the significance of different improvement strategies designed in this application.
[0119] In order to further verify the algorithm's solving performance, this application selected the six latest and best algorithms for the cross-organizational machine shop scheduling problem, namely NSGA-II, MOEA / D, DABC, KTPO, VICA, and KPMA. Each algorithm was run independently 20 times, and the average HV and GD indicators of the algorithm were statistically analyzed. The statistical results of the average HV values of DRLMCCA and the comparison algorithm are shown in Table 4, and the statistical results of the average HV values of DRLMCCA and the comparison algorithm are shown in Table 5.
[0120] Table 4
[0121]
[0122]
[0123] Table 5
[0124]
[0125]
[0126] It can be seen from Tables 4 and 5 that since the data-driven strategy of the present application can record historical search experience and also provide effective decision-making for the evolutionary computing process, the DRLMCCA algorithm of the present application has achieved significant advantages in solving this complex problem.
[0127] Furthermore, the present application can rank the Frideman rank sum test results of DRLMCCA and the comparison algorithm, as shown in Table 6.
[0128] Table 6
[0129]
[0130] As shown in Table 6, DRLMCCA achieved the highest HV and GD rankings, and the p-value was less than 0.05, which strongly demonstrated that DRLMCCA has significant advantages in comprehensive solution performance and convergence.
[0131] According to the digital-analog dual-drive intelligent scheduling method for a cross-organizational machine shop proposed in an embodiment of the present application, by modeling the actual machine shop production scenario, a cross-organizational machine shop scheduling model corresponding to the machine shop production scenario is established, and the cross-organizational machine shop scheduling model is vector-encoded to obtain the factory vector, process vector and machine vector corresponding to the cross-organizational machine shop scheduling model; the first parent population and the second parent population corresponding to the cross-organizational machine shop scheduling model are generated according to a preset collaborative initialization strategy, and a genetic evolution search operation is performed on the first parent population to obtain the first child population of the first parent population, and at the same time, a neighborhood search driven by the second parent population and preset data is performed Strategy, obtain the second child population, and record the first search knowledge corresponding to the second child population; based on the first parent population and the second child population, generate a new first parent population, and at the same time, according to the second parent population and the first child population, obtain a new second parent population, and perform a multi-operator collaborative neighborhood search operation on the new first parent population to obtain the second search knowledge corresponding to the new first parent population; horizontally splice the factory vector, process vector and machine vector according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and input the one-dimensional tensor into the pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets the preset requirements corresponding to the production scenario of the machining workshop. This application models the production scenario of the actual machining workshop as a cross-organizational machining workshop scheduling problem by considering various resource configurations and constraints, thereby improving the manufacturing capacity of aviation products in the machining workshop from the perspective of resource optimization, improving the efficiency of solving the scheduling problem of the machining workshop with complex resource grouping, and reducing the manufacturing cost of aviation products.
[0132] Secondly, the digital-analog dual-drive intelligent scheduling device for a cross-organizational machining workshop proposed according to an embodiment of the present application is described with reference to the accompanying drawings.
[0133] Fig. 9 It is a block diagram of a digital-analog dual-drive intelligent scheduling device for a cross-organizational machining workshop according to an embodiment of the present application.
[0134] like Fig. 9 As shown, the digital-analog dual-drive intelligent scheduling device 10 of the cross-organizational machining workshop includes: a modeling module 100, a data-driven module 200, a knowledge-driven module 300 and a splicing module 400.
[0135] Among them, the modeling module 100 is used to model the actual machine shop production scenario, establish a cross-organizational machine shop scheduling model corresponding to the machine shop production scenario, and vector encode the cross-organizational machine shop scheduling model to obtain the factory vector, process vector and machine vector corresponding to the cross-organizational machine shop scheduling model.
[0136] The data-driven module 200 is used to generate a first parent population and a second parent population corresponding to a cross-organizational machine shop scheduling model according to a preset collaborative initialization strategy, and perform a genetic evolutionary search operation on the first parent population to obtain a first child population of the first parent population, and at the same time obtain a second child population based on the second parent population and a preset data-driven neighborhood search strategy, and record the first search knowledge corresponding to the second child population.
[0137] The knowledge-driven module 300 is used to generate a new first parent population based on the first parent population and the second child population, and to obtain a new second parent population based on the second parent population and the first child population, and to perform a multi-operator collaborative neighborhood search operation on the new first parent population to obtain second search knowledge corresponding to the new first parent population.
[0138] The splicing module 400 is used to horizontally splice the factory vector, process vector and machine vector according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and input the one-dimensional tensor into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling plan that meets preset requirements corresponding to the production scenario of the machining workshop.
[0139] Optionally, in one embodiment of the present application, the data driving module 200 includes: a first initialization strategy unit, a second initialization strategy unit, a third initialization strategy unit, or a fourth initialization strategy unit.
[0140] The first initialization strategy unit is used to select a parallel machine with the shortest processing time for each workpiece in the cross-organizational machine shop scheduling model.
[0141] The second initialization strategy unit is used to accumulate the processing time of each workpiece in different factories to obtain the cumulative value corresponding to each workpiece, and arrange the cumulative values of all workpieces in descending order to generate a cumulative value sequence, so as to allocate each workpiece to the corresponding factory according to the cumulative value sequence.
[0142] The third initialization strategy unit is used to select the parallel machine with the minimum completion time in all factories for each job.
[0143] The fourth initialization strategy unit is used to arrange each process in ascending order according to the delivery date of each process in the cross-organizational machining workshop scheduling model to obtain a process sequence, and allocate each workpiece to a corresponding factory based on the process sequence and a preset factory code.
[0144] Optionally, in one embodiment of the present application, the data driving module 200 further includes: a calculation unit, an adjustment unit and a merging unit.
[0145] Among them, the calculation unit is used to calculate the total delay and maximum completion time corresponding to the production scenario of the computer processing workshop based on the factory vector, process vector and machine vector.
[0146] The adjustment unit is used to randomly initialize multiple target sub-populations, and adjust the factory vector, process vector and machine vector corresponding to each target sub-population in the multiple target sub-populations through any initialization strategy among multiple preset initialization strategies.
[0147] The merging unit is used to merge the multiple target sub-populations according to the adjusted factory vector, process vector and machine vector to obtain the first parent population and the second parent population.
[0148] Optionally, in one embodiment of the present application, the data driving module 200 further includes: a selection unit and a switching unit.
[0149] The selection unit is used to obtain the mating pool corresponding to the first parent population based on the two-player tournament algorithm, and randomly select a parent from the mating pool through each individual in the first parent population.
[0150] The exchange unit is used to exchange the process processing sequence of the cross-organization machine shop scheduling model according to the process-based crossover strategy and the parent, and to exchange the machine selection information and the factory information of the cross-organization machine shop scheduling model using the general crossover strategy to generate the first child population of the first parent population.
[0151] Optionally, in one embodiment of the present application, the data driving module 200 further includes: a first neighborhood search unit, a second neighborhood search unit, or a third neighborhood search unit.
[0152] Among them, the first neighborhood search unit is used to find the key workpiece with the largest delay in the cross-organizational machine shop scheduling model, and forward search for the first target workpiece that meets the preset delivery period requirements, exchange the positions of the key workpiece and the first target workpiece, or insert the key workpiece forward to the front position of the first target workpiece.
[0153] The second neighborhood search unit is used to randomly select any workpiece in the cross-organizational machine shop scheduling model, search for a second target workpiece that meets the preset delivery period requirement, and exchange the positions of the second target workpiece and any workpiece.
[0154] The third neighborhood search unit is used to find the key factory and the critical path corresponding to the key factory in the cross-organizational machine shop scheduling model according to the maximum completion time, and randomly select a key process, and scan the completion time of all machines to obtain the target machine that meets the preset completion time requirements, so as to insert the key process into the target machine.
[0155] Optionally, in one embodiment of the present application, the splicing module 400 includes: a first extraction unit, a second extraction unit, an acquisition unit, an optimization unit and an adjustment unit.
[0156] Among them, the first extraction unit is used to extract the current state of the one-dimensional tensor, and input the current state into the evaluation network of the neighborhood action recommendation neural network model to obtain the Q value of the current state, and at the same time determine the Q value of the current strategy according to the selected neighborhood action.
[0157] The second extraction unit is used to extract the next state of the one-dimensional tensor and input the next state into the target network of the neighborhood action recommendation neural network model to obtain the Q values of all actions in the next state.
[0158] The acquisition unit is used to obtain the maximum Q value of all actions in the next state based on the Q values of all actions in the next state, and calculate the target value according to the maximum Q value, a preset discount factor and the reward information of the one-dimensional tensor.
[0159] The optimization unit is used to calculate the root mean square error of the neighborhood action recommendation neural network model according to the Q value and target value of the current action, and optimize the evaluation network parameters through a preset back propagation algorithm. At the same time, when the target network parameters meet the preset update requirements, the target network parameters are updated using the optimized evaluation network parameters.
[0160] The adjustment unit is used to adjust the root mean square error through the updated target network parameters, and when the root mean square error meets the preset error condition, obtain a scheduling plan that meets the preset requirements corresponding to the production scenario of the machining workshop.
[0161] It should be noted that the aforementioned explanation of the embodiment of the digital-analog dual-drive intelligent scheduling method for cross-organizational machine shops is also applicable to the digital-analog dual-drive intelligent scheduling device for cross-organizational machine shops of this embodiment, and will not be repeated here.
[0162] According to the digital-analog dual-drive intelligent scheduling device for a cross-organizational machine shop proposed in the embodiment of the present application, it includes a modeling module, which is used to model the actual machine shop production scene, establish a cross-organizational machine shop scheduling model corresponding to the machine shop production scene, and vector encode the cross-organizational machine shop scheduling model to obtain a factory vector, a process vector and a machine vector corresponding to the cross-organizational machine shop scheduling model; a data-driven module, which is used to generate a first parent population and a second parent population corresponding to the cross-organizational machine shop scheduling model according to a preset collaborative initialization strategy, and perform a genetic evolution search operation on the first parent population to obtain a first child population of the first parent population, and at the same time perform a neighborhood search based on the second parent population and the preset data-driven Strategy, obtain the second child population, and record the first search knowledge corresponding to the second child population; knowledge-driven module, for generating a new first parent population based on the first parent population and the second child population, and obtaining a new second parent population based on the second parent population and the first child population, and performing a multi-operator collaborative neighborhood search operation on the new first parent population to obtain the second search knowledge corresponding to the new first parent population; splicing module, for horizontally splicing the factory vector, process vector and machine vector according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and input the one-dimensional tensor into the pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets the preset requirements corresponding to the production scenario of the machining workshop. This application models the production scenario of the actual machining workshop as a cross-organizational machining workshop scheduling problem by considering various resource configurations and constraints, thereby improving the manufacturing capacity of aviation products in the machining workshop from the perspective of resource optimization, improving the efficiency of solving the scheduling problem of the machining workshop with complex resource grouping, and reducing the manufacturing cost of aviation products.
[0163] Fig.10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:
[0164] A memory 1001 , a processor 1002 , and a computer program stored in the memory 1001 and executable on the processor 1002 .
[0165] When the processor 1002 executes the program, the digital-analog dual-drive intelligent scheduling method for cross-organizational machining workshops provided in the above embodiment is implemented.
[0166] Furthermore, the electronic device further comprises:
[0167] The communication interface 1003 is used for communication between the memory 1001 and the processor 1002 .
[0168] The memory 1001 is used to store computer programs that can be executed on the processor 1002 .
[0169] The memory 1001 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
[0170] If the memory 1001, the processor 1002 and the communication interface 1003 are implemented independently, the communication interface 1003, the memory 1001 and the processor 1002 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.10 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.
[0171] Optionally, in a specific implementation, if the memory 1001, the processor 1002 and the communication interface 1003 are integrated on a chip, the memory 1001, the processor 1002 and the communication interface 1003 can communicate with each other through an internal interface.
[0172] The processor 1002 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.
[0173] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned digital-analog dual-drive intelligent scheduling method for cross-organizational machining workshops.
[0174] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.
[0175] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0176] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.
[0177] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.
[0178] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0179] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.
[0180] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
[0181] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.
Claims
1. A digital-analog dual-drive intelligent scheduling method for cross-organizational machining workshops, characterized in that: The following steps are involved: Modeling an actual machining workshop production scenario, establishing a cross-organizational machining workshop scheduling model corresponding to the machining workshop production scenario, and vector encoding the cross-organizational machining workshop scheduling model to obtain a factory vector, a process vector, and a machine vector corresponding to the cross-organizational machining workshop scheduling model; Generate a first parent population and a second parent population corresponding to the cross-organizational machine shop scheduling model according to a preset collaborative initialization strategy, perform a genetic evolution search operation on the first parent population to obtain a first child population of the first parent population, and simultaneously obtain a second child population based on the second parent population and a preset data-driven neighborhood search strategy, and record first search knowledge corresponding to the second child population; Based on the first parent population and the second child population, a new first parent population is generated, and based on the second parent population and the first child population, a new second parent population is obtained, and a multi-operator collaborative neighborhood search operation is performed on the new first parent population to obtain second search knowledge corresponding to the new first parent population; The factory vector, the process vector and the machine vector are horizontally spliced according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and the one-dimensional tensor is input into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling plan that meets preset requirements corresponding to the production scenario of the machining workshop.
2. The method according to claim 1, characterized in that The preset collaborative initialization strategy includes: Selecting a parallel machine with the smallest processing time for each workpiece in the cross-organizational machine shop scheduling model; Alternatively, the processing time of each workpiece in different factories is accumulated to obtain the accumulated value corresponding to each workpiece, and the accumulated values of all workpieces are arranged in descending order to generate an accumulated value sequence, so as to allocate each workpiece to a corresponding factory according to the accumulated value sequence; Alternatively, each of the workpieces selects a parallel machine with the minimum completion time in all factories; Alternatively, according to the delivery date of each process in the cross-organizational machine shop scheduling model, each process is arranged in ascending order to obtain a process sequence, and based on the process sequence and a preset factory code, each workpiece is allocated to a corresponding factory.
3. The method according to claim 1, characterized in that: The generating of the first parent population and the second parent population corresponding to the cross-organizational machine shop scheduling model according to the preset collaborative initialization strategy includes: Calculate the total delay and the maximum completion time corresponding to the production scenario of the machining workshop according to the factory vector, the process vector and the machine vector; Initializing a plurality of target sub-populations randomly, and adjusting the factory vector, the process vector, and the machine vector corresponding to each target sub-population in the plurality of target sub-populations by any one of a plurality of preset initialization strategies; The multiple target sub-populations are merged according to the adjusted factory vector, the process vector, and the machine vector to obtain the first parent population and the second parent population.
4. The method according to claim 3, characterized in that The performing a genetic evolution search operation on the first parent population to obtain a first child population of the first parent population includes: Based on a two-player tournament algorithm, obtaining a mating pool corresponding to the first parent population, and randomly selecting a parent from the mating pool through each individual in the first parent population; The process processing sequence of the cross-organizational machine shop scheduling model is exchanged with the parent according to the process-based crossover strategy, and the machine selection information and factory information of the cross-organizational machine shop scheduling model are exchanged using a general crossover strategy to generate a first child population of the first parent population.
5. The method according to claim 4, characterized in that The preset data-driven neighborhood search strategy includes: Find the key workpiece with the largest delay in the cross-organizational machine shop scheduling model, and search forward for a first target workpiece that meets the preset delivery date requirement, and exchange the positions of the key workpiece and the first target workpiece, or insert the key workpiece forward to the position in front of the first target workpiece; Alternatively, any workpiece in the cross-organizational machine shop scheduling model is randomly selected, and a second target workpiece that meets the preset delivery date requirement is searched, and the positions of the second target workpiece and any workpiece are exchanged; Alternatively, the key factory and the critical path corresponding to the key factory of the cross-organizational machine shop scheduling model are searched according to the maximum completion time, and a key process is randomly selected. The completion time of all machines is scanned to obtain a target machine that meets the preset completion time requirement, so as to insert the key process into the target machine.
6. The method according to claim 5, characterized in that The one-dimensional tensor is input into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling solution that meets preset requirements corresponding to the production scenario of the machining workshop, including: Extracting the current state of the one-dimensional tensor, and inputting the current state into the evaluation network of the neighborhood action recommendation neural network model to obtain the Q value of the current state, and determining the Q value of the current strategy according to the selected neighborhood action; Extracting the next state of the one-dimensional tensor, and inputting the next state into the target network of the neighborhood action recommendation neural network model to obtain the Q values of all actions in the next state; Based on the Q values of all actions in the next state, obtain the maximum Q value among all actions in the next state, and calculate the target value according to the maximum Q value, a preset discount factor and the reward information of the one-dimensional tensor; Calculating the root mean square error of the neighborhood action recommendation neural network model according to the Q value of the current action and the target value, optimizing the evaluation network parameters through a preset back propagation algorithm, and updating the target network parameters using the optimized evaluation network parameters when the target network parameters meet the preset update requirements; The root mean square error is adjusted through the updated target network parameters, and when the root mean square error meets the preset error condition, a scheduling plan that meets the preset requirements corresponding to the production scenario of the machining workshop is obtained.
7. A digital-analog dual-drive intelligent scheduling device for cross-organizational machining workshops, characterized in that: include: A modeling module, used to model the actual machine shop production scenario, establish a cross-organization machine shop scheduling model corresponding to the machine shop production scenario, and vector encode the cross-organization machine shop scheduling model to obtain a factory vector, a process vector, and a machine vector corresponding to the cross-organization machine shop scheduling model; A data-driven module, used to generate a first parent population and a second parent population corresponding to the cross-organizational machine shop scheduling model according to a preset collaborative initialization strategy, and perform a genetic evolution search operation on the first parent population to obtain a first child population of the first parent population, and simultaneously obtain a second child population based on the second parent population and a preset data-driven neighborhood search strategy, and record the first search knowledge corresponding to the second child population; A knowledge-driven module, configured to generate a new first parent population based on the first parent population and the second child population, obtain a new second parent population based on the second parent population and the first child population, and perform a multi-operator collaborative neighborhood search operation on the new first parent population to obtain second search knowledge corresponding to the new first parent population; A splicing module is used to horizontally splice the factory vector, the process vector and the machine vector according to the first search knowledge and the second search knowledge to obtain a one-dimensional tensor, and input the one-dimensional tensor into a pre-trained neighborhood action recommendation neural network model to obtain a scheduling plan that meets preset requirements corresponding to the production scenario of the machining workshop.
8. The device according to claim 7, characterized in that The data driving module comprises: A first initialization strategy unit is used to select a parallel machine with the shortest processing time for each workpiece in the cross-organizational machine shop scheduling model; or a second initialization strategy unit, configured to accumulate the processing time of each workpiece in different factories to obtain a cumulative value corresponding to each workpiece, and to arrange the cumulative values of all workpieces in descending order to generate a cumulative value sequence, so as to allocate each workpiece to a corresponding factory according to the cumulative value sequence; or A third initialization strategy unit is used for selecting, for each workpiece, a parallel machine with the minimum completion time in all factories; or, The fourth initialization strategy unit is used to arrange each process in ascending order according to the delivery date of each process in the cross-organizational machining workshop scheduling model to obtain a process sequence, and allocate each workpiece to a corresponding factory based on the process sequence and a preset factory code.
9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the digital-analog dual-drive intelligent scheduling method for cross-organizational machine shops as described in any one of claims 1 to 6.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the digital-analog dual-drive intelligent scheduling method for a cross-organizational machining workshop as described in any one of claims 1-6.
Citation Information
Patent Citations
Operation workshop scheduling modeling method based on genetic algorithm
CN103870647A
Production scheduling method and system based on hybrid parallel inheritance and variable neighborhood algorithm
CN113705812A
Flexible job shop scheduling method and system based on hybrid evolutionary algorithm
CN114580743A
Multi-park fuzzy blanking workshop system and method
CN116151424A
Method and device for scheduling job shop
JP1996315028A
Cited By
Intelligent scheduling system and method for multi-station numerical control machining equipment
CN120941074A
