Dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm
By adopting auxiliary task collaborative optimization algorithm and real-time detection framework for order changes in flexible operation workshops, the problems of order cancellation and emergency order insertion are solved, dynamic response and scheduling optimization for order changes are achieved, and the flexibility of workshop scheduling is improved Sex and efficiency.
Patent Information
- Application Number
- CN202411021921.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-12
AI Technical Summary
The prior art is difficult to effectively deal with the situation where order cancellation and emergency order insertion exist in flexible operation workshops, which leads to difficulty in optimizing the scheduling plan and lacks a response mechanism for order uncertainty and emergencies.
Using dynamic flexible work workshop scheduling method based on auxiliary task collaborative optimization algorithm, a real-time detection framework for order changes is built, and a rescheduling scheme is generated. Combining the collaborative optimization algorithm and task optimization mathematical model, the workshop scheduling scheme is optimized to deal with order insertion, cancellation and insertion The complex situation of disappearing.
It realizes timely response to order changes, improves the feasibility and convergence of the scheduling plan, and can obtain a good optimization scheduling plan in the multi-task optimization process, which is suitable for the dynamic scheduling needs of the intelligent manufacturing workshop.
Smart Images

Figure CN118941039A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of workshop production scheduling, and in particular to a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm. Background Art
[0002] Workshop production scheduling is the foundation of the manufacturing system, and the optimization of production scheduling is the core of advanced manufacturing technology and modern management technology. The survival and competition of manufacturing enterprises depend on whether they can produce customized, multi-variety, and small-batch products in a short production cycle. With the emergence of flexible production systems such as flexible manufacturing systems and CNC machining centers, the research results of classic job shop scheduling cannot be directly applied to manufacturing systems with flexible characteristics. Therefore, the problem of flexible job shop scheduling (FJSP) has become the focus of current research. In flexible job shop scheduling, each process of the workpiece can be processed on multiple machines, and the processing time on different machines may not be the same. Scientifically formulating production scheduling plans plays a vital role in shortening product production cycles, controlling workshop work-in-process inventory, improving product delivery date satisfaction rates, and improving enterprise productivity.
[0003] At present, most studies consider the workshop scheduling problem in a static environment. In practical applications, order changes in the workshop scheduling problem are usually random and accompanied by random constraints. In response to order changes, some researchers have studied the rescheduling after the insertion of emergency orders, and some researchers have studied the situation of order cancellation.
[0004] In addition, traditional genetic algorithms have the problem of "premature" convergence. The algorithm will no longer evolve due to being trapped in the local optimal solution, which will affect the evolutionary optimization of the algorithm. The multi-task evolutionary optimization algorithm is to study the simultaneous solution of multiple optimization tasks to independently improve the performance of solving each task. Its working principle is that there is some common useful knowledge when solving a certain task. The useful knowledge obtained in the process of solving the task may help solve another task related to it. Therefore, in the workshop scheduling problem, there are multiple optimization tasks, such as minimum completion time, minimum maximum machine load, minimum overall machine load, minimum workpiece deviation, etc., and different tasks are related. The ultimate goal of multiple tasks is to meet the robustness and high-efficiency production requirements of the manufacturing system. Therefore, there is an urgent need for an optimization algorithm to solve the problem of flexible job shop scheduling with uncertain orders and realize efficient production of intelligent manufacturing in job shops.
[0005] However, in actual production, changes in orders may involve the coexistence of these two situations. The uncertainty of emergencies faced by uncertain orders brings more challenges to the optimization of scheduling plans. Currently, there is no technology to study the situation where order cancellations and emergency orders exist at the same time.
[0006] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention
[0007] 1. Technical issues to be resolved
[0008] In view of the shortcomings of the prior art, the present invention provides a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm, which has the advantage of timely responding to dynamic events of order changes occurring in the actual manufacturing process, thereby solving the problem in the prior art that the situation in which order cancellation and emergency order insertion exist at the same time is not handled.
[0009] (II) Technical solution
[0010] In order to achieve the above-mentioned advantage of timely responding to dynamic events of order changes occurring in the actual manufacturing process, the specific technical solution adopted by the present invention is as follows:
[0011] A dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm, the dynamic flexible job shop scheduling method comprises the following steps:
[0012] S1. Build a real-time order change detection framework for dynamic flexible job shop scheduling, and design a rescheduling plan based on the detection results and the real-time processing process of the workshop;
[0013] S2. Generate a target optimization function based on the rescheduling plan and the workshop optimization task objectives, and build a task optimization mathematical model based on the target optimization function and the workshop conditions;
[0014] S3. Based on the auxiliary tasks in the dynamic flexible job shop, the collaborative optimization algorithm is combined with the task optimization mathematical model to optimize the rescheduling plan to obtain the final rescheduling plan in the dynamic flexible job shop.
[0015] Preferably, in view of the order changes in the dynamic flexible job shop scheduling process, a real-time order change detection framework is constructed, and a rescheduling plan is designed based on the detection results and the real-time processing process of the workshop, including the following steps:
[0016] S11. Execute the pre-scheduling plan at the beginning of the dynamic flexible job shop scheduling, and build a real-time detection framework for order changes during the execution process, and use the detection framework to monitor the execution process of the pre-scheduling plan;
[0017] S12, judging whether the order has changed dynamically based on the monitoring result, if the order has not changed dynamically, continuing to execute the pre-scheduling plan, if the order has changed dynamically, executing step S13;
[0018] S13. Determine the corresponding order change type based on the dynamic change status of the order, and select a rescheduling strategy to generate a rescheduling plan according to the order change type;
[0019] S14. Execute the rescheduling plan, use the real-time order change detection framework to monitor the order status after the rescheduling plan is executed, and determine whether the order has changed dynamically based on the monitoring results.
[0020] Preferably, judging the corresponding order change type based on the dynamic change status of the order, and selecting the rescheduling strategy to generate the rescheduling plan according to the order change type includes the following steps:
[0021] S131. Obtain the change type of the corresponding order according to the dynamic change status of the order, where the change type includes order insertion, order cancellation, and order insertion and cancellation coexisting;
[0022] S132. When the order change type is order cancellation, cancel the time node of the corresponding order and reschedule all unprocessed workpieces;
[0023] S133. When the order change type is order insertion, reschedule the unfinished orders based on the pre-scheduling plan, select a machine for the new order and determine the machine startup time;
[0024] S134. When the order change type is that order insertion and cancellation exist at the same time, the order scheduling operation is performed based on the workpiece floating time insertion rescheduling strategy.
[0025] Preferably, when the order change type is that order insertion and cancellation exist at the same time, the order scheduling operation based on the workpiece floating time insertion rescheduling strategy includes the following steps:
[0026] S1341. When the order change type is order insertion and cancellation coexisting, determine the floating time of the machine corresponding to the new order according to the pre-scheduling plan. The floating time includes the available idle time, total idle time and gap idle time of the corresponding machine.
[0027] S1342. If the idle time of the new order is less than or equal to the available idle time, the new order is directly assigned to the corresponding machine position;
[0028] S1343. If the idle time of the new order is greater than the available idle time and less than or equal to the sum of the available idle time and the gap idle time, the new order is assigned to the corresponding machine and the preset process of the corresponding machine is shifted right;
[0029] S1344. If the idle time of the new order is greater than the sum of the available idle time and the gap idle time, and less than or equal to the sum of the available idle time and the total idle time, the new order is assigned to the corresponding machine, and the positions of the corresponding workpieces and machines are adjusted;
[0030] S1345. If the idle time of the new order does not meet all the idle time intervals of the corresponding machines, the standby machine is enabled and the order operation is performed in the order of the new order. If the idle time of the new order does not meet the idle time intervals on all the machines, the process of the new order is placed at the end of the pre-scheduling plan.
[0031] S1346. Repeat steps S1342 to S1345 to arrange all new orders. If all new orders are arranged, the order arrangement operation is performed. If there are new orders that have not been arranged, repeat steps S1342 to S1345.
[0032] Preferably, the calculation formula corresponding to the machine gap idle time is:
[0033]
[0034]
[0035]
[0036] In the formula, Indicates the gap idle time, Indicates the current process O i,j The end time and next process O i,j+1 The difference between the start times, Indicates the current process O i,j The completion time of the current process O i,j Next process O on the machine i,j+1 The difference between the start times, S i,j+1 Indicates the next process O i,j+1 The time to start processing, C ij Indicates the current process O i,j End processing time, r i represents the total number of processes for job i, j represents the process index, C ect Indicates the current process O i,j Estimated completion time, Ci,j represents the current process O i,j The end processing time, m represents the total number of machines in the dynamic flexible job shop, k represents the machine index, and h k represents the number of processes that need to be arranged on the machine, P represents the index of the process position on the machine, S k,P+1 represents the start time of the process at position P+1 on machine k, Ck,P represents the end time of the process at position P on machine k, Y i,j,k,P Indicates process o ij Processing time at position P of machine k, C k,hk represents position h on machine k k The process completion time, Y i,j,k,hk Indicates process o ij At position h on machine k k Processing time.
[0037] Preferably, generating a target optimization function based on the rescheduling plan and the workshop optimization task target, and constructing a task optimization mathematical model according to the target optimization function and the workshop status includes the following steps:
[0038] S21, obtaining the completion time, machine load information and workpiece flow time after the order is dynamically changed according to the rescheduling plan, and obtaining the completion time, machine load information and workpiece flow time of the pre-scheduling plan;
[0039] S22, combining the completion time, machine load information and overall workpiece flow time with the workshop optimization task target to generate a target optimization function;
[0040] S23. Generate constraint conditions according to the target optimization function, and construct a task optimization mathematical model based on the target optimization function, the constraint conditions and the workshop conditions.
[0041] Preferably, combining the completion time, machine load information and overall workpiece flow time with the workshop optimization task objectives to generate a target optimization function includes the following steps:
[0042] S221, generating a time efficiency objective function and a machine robustness objective function according to the completion times corresponding to the rescheduling plan and the pre-scheduling plan;
[0043] S222, generating a maximum machine load deviation objective function according to the machine load information corresponding to the rescheduling scheme and the prescheduling scheme;
[0044] S223, generating an overall workpiece flow time deviation objective function using the overall workpiece flow time corresponding to the rescheduling scheme and the prescheduling scheme;
[0045] S224, generating a dual-objective optimization function based on the time efficiency objective function, the machine robustness objective function, the maximum machine load deviation objective function and the overall workpiece flow time deviation objective function.
[0046] Preferably, the expression of the dual-objective optimization function is:
[0047]
[0048] Where f1 represents the time efficiency objective function, f2 represents the machine robustness objective function, f3 represents the maximum machine load deviation objective function, f4 represents the overall workpiece flow time deviation objective function, and x represents the decision variable.
[0049] Preferably, based on the auxiliary tasks in the dynamic flexible job shop, the collaborative optimization algorithm is combined with the task optimization mathematical model to optimize the rescheduling scheme to obtain the final rescheduling scheme in the dynamic flexible job shop, including the following steps:
[0050] S31. Generate an initial population of task optimization based on the pre-scheduling plan, and divide the tasks into main tasks and auxiliary tasks according to the job requirements of the dynamic flexible job shop;
[0051] S32. Establish a main task optimization probability model based on the collaborative optimization algorithm, and use the main task optimization probability model to determine whether there is cross-task knowledge transfer;
[0052] S33. If there is cross-task knowledge transfer, individuals in the auxiliary task are transferred to the main task based on the task knowledge transfer strategy to perform knowledge fusion operation;
[0053] S34. The dynamic change results of the orders after the execution of the rescheduling plan are combined with the process floating time and the rescheduling decoding strategy inserted according to the task optimization mathematical model, and the rescheduling plan is decoded to obtain the final rescheduling plan of the dynamic flexible job shop.
[0054] Preferably, the task optimization initial population includes a main population and an auxiliary population;
[0055] The main population is the process with the largest remaining number of processes and follows the rule of minimum machine workload allocation; the auxiliary population is the process that follows the rule of global minimum processing time.
[0056] (III) Beneficial effects
[0057] Compared with the prior art, the present invention provides a dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm, which has the following beneficial effects:
[0058] (1) The present invention takes into account the dynamic nature of order changes in actual production during workshop scheduling, constructs a real-time detection framework with uncertain dynamic order changes, and is able to respond promptly to dynamic events of order changes occurring during the actual manufacturing process. Compared with the prior art, the present invention is more suitable for handling scheduling problems in smart manufacturing workshops in actual production.
[0059] (2) In the face of different types of real-time changes in orders, the present invention designs a dynamic response mechanism to generate a rescheduling plan. In the face of different types of order changes, different dynamic responses are used to generate rescheduling plans, which accelerates the response time for handling dynamic events and improves the feasibility of the rescheduling plan. At the same time, based on the diversity characteristics of the collaborative optimization of multiple tasks in workshop scheduling, the auxiliary task collaborative optimization algorithm is used to solve the dynamic scheduling solution, thereby improving the algorithm's solution capabilities in terms of convergence and diversity, so that a good optimized scheduling plan can be obtained in the optimization process of multiple tasks.
[0060] (3) Inspired by the multi-task evolutionary optimization algorithm, the present invention can optimize multiple different tasks at the same time. When the main task search falls into the local optimum, the auxiliary task provides diverse solutions for the target task and finally obtains the optimal scheduling solution. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0062] Figure 1 is a flow chart of a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0063] Figure 2 is a schematic diagram of a scheduling scheme detection framework in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0064] Figure 3 It is a schematic diagram of a framework of an auxiliary task collaborative optimization algorithm in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0065] Figure 4 is a schematic diagram of a population initialization area in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0066] Figure 5 is a schematic diagram of a crossover method in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0067] Figure 6 It is one of the schematic diagrams of IGD values of five algorithms running each time on the DMK benchmark instance in the dynamic flexible job shop scheduling method based on the auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0068] Figure 7 This is the second schematic diagram of IGD values of five algorithms running each time on a DMK benchmark instance in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0069] Figure 8 This is a third schematic diagram of IGD values of five algorithms running each time on a DMK benchmark instance in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0070] Fig. 9 This is a fourth schematic diagram of IGD values of five algorithms running each time on a DMK benchmark instance in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0071] Fig.10 It is one of the schematic diagrams of MS values of five algorithms running each time on the DMK benchmark instance in the dynamic flexible job shop scheduling method based on the auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0072] Fig.11 This is the second schematic diagram of MS values of five algorithms running each time on a DMK benchmark instance in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0073] Fig.12 This is a third schematic diagram of MS values of five algorithms running each time on a DMK benchmark instance in a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm according to an embodiment of the present invention;
[0074] Fig.13 This is the fourth schematic diagram of MS values of five algorithms running each time on the DMK benchmark instance in the dynamic flexible job shop scheduling method based on the auxiliary task collaborative optimization algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION
[0075] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments. They can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in this field should be able to understand other possible implementation methods and advantages of the present invention.
[0076] According to an embodiment of the present invention, a dynamic flexible job shop scheduling method based on an auxiliary task collaborative optimization algorithm is provided.
[0077] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1-Figure 5As shown, according to the dynamic flexible job shop scheduling method based on the auxiliary task collaborative optimization algorithm according to an embodiment of the present invention, the dynamic flexible job shop scheduling method includes the following steps:
[0078] S1. Aiming at the order changes in the dynamic flexible job shop scheduling process, a real-time order change detection framework is constructed, and a rescheduling plan is designed based on the detection results and the real-time processing process of the workshop.
[0079] In this embodiment, a real-time order change detection framework is constructed for order changes in the dynamic flexible job shop scheduling process, and a rescheduling plan is designed based on the detection results and the real-time processing process of the workshop, including the following steps:
[0080] S11. Execute the pre-scheduling plan at the beginning of the dynamic flexible job shop scheduling, and build a real-time detection framework for order changes during the execution process, and use the detection framework to monitor the execution process of the pre-scheduling plan;
[0081] S12, judging whether the order has changed dynamically based on the monitoring result, if the order has not changed dynamically, continuing to execute the pre-scheduling plan, if the order has changed dynamically, executing step S13;
[0082] S13. Determine the corresponding order change type based on the dynamic change status of the order, and select a rescheduling strategy to generate a rescheduling plan according to the order change type;
[0083] S14. Execute the rescheduling plan, use the real-time order change detection framework to monitor the order status after the rescheduling plan is executed, and determine whether the order has changed dynamically based on the monitoring results.
[0084] Specifically, judging the corresponding order change type based on the dynamic change status of the order, and selecting the rescheduling strategy to generate the rescheduling plan according to the order change type includes the following steps:
[0085] S131. Obtain the change type of the corresponding order according to the dynamic change status of the order, where the change type includes order insertion, order cancellation, and order insertion and cancellation coexisting;
[0086] S132. When the order change type is order cancellation, cancel the time node of the corresponding order and reschedule all unprocessed workpieces;
[0087] S133. When the order change type is order insertion, reschedule the unfinished orders based on the pre-scheduling plan, select a machine for the new order and determine the machine startup time;
[0088] S134. When the order change type is that order insertion and cancellation exist at the same time, the order scheduling operation is performed based on the workpiece floating time insertion rescheduling strategy.
[0089] It needs to be explained that when a new order is inserted, only the new order workpiece is scheduled. On the basis of the original scheduling plan, the machine selection and start time of the new order are determined. This means that when scheduling a new order, each machine has a different start time, and the earliest start time of the new order is not zero.
[0090] At the same time, after the new order is inserted, the new order artifacts and unfinished order artifacts are rearranged. When the new order insertion instruction is received, some jobs have been processed, and some jobs have not yet been processed. All unprocessed artifacts are rearranged. This strategy means that all jobs and machines have different start times, and the scheduling problem is transformed into FJSP (flexible job shop scheduling) with different machine and job start times.
[0091] The method of crossing orders from different areas is adopted. In the crossing method, the precedence operation crossover (POX) based on priority order is used in the process chain. When facing order cancellation, the area between L1 and L2 is called crossover area 1, the area between L2 and all final operations is called crossover area 2, and all operation intervals after L1 are called crossover area 3. When facing order cancellation and insertion, the position between L1 and L2 is called crossover area 1, the position between L3 and L4 is called crossover area L2, and all operation intervals after L1 are called cross-area 3.
[0092] Process crossover mode 1: POX (crossover area 1) + POX (crossover area 2);
[0093] Process crossover method 2: POX (crossover area 3).
[0094] Among them, when the order change type is that order insertion and cancellation exist at the same time, the order scheduling operation based on the workpiece floating time insertion rescheduling strategy includes the following steps:
[0095] S1341. When the order change type is order insertion and cancellation, determine the floating time of the machine corresponding to the new order according to the pre-scheduling plan. The floating time includes the available idle time of the corresponding machine, the total idle time (T I ) and intermittent idle time (Partial idle time, P I );
[0096] S1342. If the idle time of the new order is less than or equal to the available idle time, the new order is directly assigned to the corresponding machine position;
[0097] S1343. If the idle time of the new order is greater than the available idle time and less than or equal to the sum of the available idle time and the gap idle time, the new order is assigned to the corresponding machine and the preset process of the corresponding machine is shifted right;
[0098] S1344. If the idle time of the new order is greater than the sum of the available idle time and the gap idle time, and less than or equal to the sum of the available idle time and the total idle time, the new order is assigned to the corresponding machine, and the positions of the corresponding workpieces and machines are adjusted;
[0099] S1345. If the idle time of the new order does not meet all the idle time intervals of the corresponding machines, the standby machine is enabled and the order operation is performed in the order of the new order. If the idle time of the new order does not meet the idle time intervals on all the machines, the process of the new order is placed at the end of the pre-scheduling plan.
[0100] S1346. Repeat steps S1342 to S1345 to arrange all new orders. If all new orders are arranged, the order arrangement operation is performed. If there are new orders that have not been arranged, repeat steps S1342 to S1345.
[0101] At the same time, the calculation formula for the corresponding machine gap idle time is:
[0102]
[0103]
[0104]
[0105] In the formula, Indicates the gap idle time, Indicates the current process O i,j The end time and next process O i,j+1 The difference between the start times, Indicates the current process O i,j The completion time of the current process O i,j Next process O on the machine i,j+1 The difference between the start times, S i,j+1 Indicates the next process O i,j+1 The time to start processing, C ij Indicates the current process O i,j End processing time, r i represents the total number of processes for job i, j represents the process index, C i,j Indicates the current process O i,j The end processing time, m represents the total number of machines in the dynamic flexible job shop, k represents the machine index, and h krepresents the number of processes that need to be arranged on the machine, P represents the index of the process position on the machine, S k,P+1 represents the start time of the process at position P+1 on machine k, C k,P represents the end time of the process at position P on machine k, Y i,j,k,P Indicates process o ij The value of processing at position P of machine k is 1, otherwise it is 0, C k,hk represents position h on machine k k The process completion time, Y i,j,k,hk Indicates process o ij At position h on machine k k 1 if it is processed, otherwise 0, C ect Indicates the current process O i,j Estimated time of completion.
[0106] If there is a subsequent job, the difference between the start time of the subsequent workpiece and the completion time of the current workpiece is the partial float time of the workpiece. If there is no subsequent workpiece, the difference between the reference completion time and the completion time of the current workpiece is regarded as the partial float time.
[0107] S2. Generate a target optimization function based on the rescheduling plan and the workshop optimization task objectives, and build a task optimization mathematical model based on the target optimization function and the workshop conditions.
[0108] It should be explained that the dynamic scheduling scheme forms a closed-loop optimization in the production process through real-time detection and monitoring, and then the following conditions need to be met when establishing a mathematical model for multi-objective flexible job shop scheduling with dynamic order changes:
[0109]
[0110]
[0111]
[0112]
[0113]
[0114]
[0115]
[0116]
[0117]
[0118]
[0119]
[0120] Specifically, inequality (1) and inequality (2) are constraints on the operation sequence of the workpiece, equation (3) is used to calculate the completion time of the overall scheduling plan, inequality (4) is the constraint on the completion time of each process, the completion time of each process cannot exceed the completion time of the overall plan, equation (5) is used to calculate the machine load, inequality (6) and inequality (7) are constraints on the ability of the same machine to process only one process at the same time, equations (8) and (9) indicate that each machine has a cycle operation, equation (10) is a constraint on the machine, the same process can only be processed by one machine, and inequality (11) indicates that each parameter variable is a positive number.
[0121] Where i and u represent workpiece indexes, j and v represent process indexes, and j = 1, 2, ..., n i , u=1,2,…,n u , k represents the machine index, k = 1, 2, ..., m; n represents the total number of workpieces in the workshop, r i represents the total number of processes for workpiece i, m represents the total number of machines in the workshop, and o ij represents the jth process of workpiece i, T i,j,k Indicates process o ij Processing time on machine k, S ij Indicates process o ij Start processing time, C ij Indicates process o ij End processing time, P i,j,k Indicates process o ij Processing time on machine k, C max represents the completion time of the scheduling plan, X i,j,k If process o ij Processed on machine k, X i,j,k =1; otherwise X i,j,k =0, b i,j,k,u,v It means that if on machine k, process O i,j In process O u,v Before being processed, b i,j,k,u,v =1; otherwise, b i,j,k,u,v =0.
[0122] In this embodiment, generating a target optimization function based on the rescheduling plan and the workshop optimization task target, and constructing a task optimization mathematical model according to the target optimization function and the workshop status includes the following steps:
[0123] S21, obtaining the completion time, machine load information and workpiece flow time after the order is dynamically changed according to the rescheduling plan, and obtaining the completion time, machine load information and workpiece flow time of the pre-scheduling plan;
[0124] S22, combining the completion time, machine load information and overall workpiece flow time with the workshop optimization task target to generate a target optimization function;
[0125] S23. Generate constraint conditions according to the target optimization function, and construct a task optimization mathematical model based on the target optimization function, the constraint conditions and the workshop conditions.
[0126] Specifically, the completion time, machine load information, and overall workpiece flow time are combined with the workshop optimization task objectives to generate a target optimization function, including the following steps:
[0127] S221, generating a time efficiency objective function and a machine robustness objective function according to the completion times corresponding to the rescheduling plan and the pre-scheduling plan;
[0128] S222, generating a maximum machine load deviation objective function according to the machine load information corresponding to the rescheduling scheme and the prescheduling scheme;
[0129] S223, generating an overall workpiece flow time deviation objective function using the overall workpiece flow time corresponding to the rescheduling scheme and the prescheduling scheme;
[0130] S224, generating a dual-objective optimization function based on the time efficiency objective function, the machine robustness objective function, the maximum machine load deviation objective function and the overall workpiece flow time deviation objective function.
[0131] It needs to be explained that, based on the uncertainty of order changes, the expression of the multi-task optimization mathematical model is established:
[0132]
[0133] Decision variable x = (x1, x2, x D ), D is the dimension of the decision variable, the objective function is F→Ω: composed of multiple objectives, M is the number of objective functions, and the mathematical model includes the following multiple optimization objective functions (dual objective optimization function):
[0134]
[0135] Where f1 represents the time efficiency objective function, f2 represents the machine robustness objective function, f3 represents the maximum machine load deviation objective function, f4 represents the overall workpiece flow time deviation objective function, and x represents the decision variable.
[0136] F1-F3 are three two-objective optimization functions, f1-f4 are four single-objective optimization functions, and the mathematical expressions of f1-f4 are as follows:
[0137]
[0138] f2=SS=min(J deviation );
[0139]
[0140]
[0141] In the formula, is the actual completion time after dynamic changes, is the estimated completion time in the original scheduling plan, is the actual machine load after dynamic changes, is the machine load in the pre-scheduling scheme, is the overall workpiece flow time after dynamic changes, is the overall job flow time in the pre-scheduling solution.
[0142] Three two-objective functions are considered, F1(x) is a two-objective function of time efficiency (TE) and robustness (solution stability, SS), which is the objective function of the main optimization task in the design algorithm, F2(x) is a two-objective function of time efficiency and maximum machine workload deviation (MMWD), which is the objective function for auxiliary task optimization, and F3(x) is a two-objective function of solution stability and total jobs flow deviation (TJFD), which is the objective function for auxiliary task optimization.
[0143] Therefore, the mathematical model description of dynamic order change multi-objective flexible job shop scheduling is as follows:
[0144]
[0145]
[0147]
[0148]
[0149] In the formula, C i represents the processing completion time of the i-th workpiece, C r Oij represents the process completion time of the pre-scheduling plan, C p Oijrepresents the process completion time of the rescheduling plan, n represents the total number of workpieces in the workshop, n' represents the total number of affected workpieces in the workshop after the dynamic event occurs, q' represents the total number of affected processes in the workshop after the dynamic event occurs, i represents the workpiece index, j represents the process index, and k represents the machine index.
[0150] S3. Based on the auxiliary tasks in the dynamic flexible job workshop, the collaborative optimization algorithm is combined with the task optimization mathematical model to optimize the rescheduling plan to obtain the final rescheduling plan in the dynamic flexible job workshop.
[0151] In this embodiment, based on the auxiliary tasks in the dynamic flexible job shop, the collaborative optimization algorithm and the task optimization mathematical model are combined to optimize the rescheduling scheme to obtain the final rescheduling scheme in the dynamic flexible job shop, including the following steps:
[0152] S31. Generate an initial population of task optimization based on the pre-scheduling plan, and divide the tasks into main tasks and auxiliary tasks according to the job requirements of the dynamic flexible job shop;
[0153] S32. Establish a main task optimization probability model based on the collaborative optimization algorithm, and use the main task optimization probability model to determine whether there is cross-task knowledge transfer;
[0154] S33. If there is cross-task knowledge transfer, individuals in the auxiliary task are transferred to the main task based on the task knowledge transfer strategy to perform knowledge fusion operation;
[0155] S34. The dynamic change results of the orders after the execution of the rescheduling plan are combined with the process floating time and the rescheduling decoding strategy inserted according to the task optimization mathematical model, and the rescheduling plan is decoded to obtain the final rescheduling plan of the dynamic flexible job shop.
[0156] Specifically, the initial population for task optimization includes the main population and the auxiliary population;
[0157] The main population is the process with the largest number of remaining processes and follows the rule of minimum machine allocation workload; the auxiliary population is the process that follows the rule of global minimum processing time, and the auxiliary population is only reinitialized within the affected interval 1, and the process with the largest number of remaining processes is preferentially selected. For the main population, only the jobs within the affected interval 2 are initialized, and the initialization method is a random rule;
[0158] For the auxiliary population: the global minimum processing time rule is used for machine allocation, and for the primary population: the workload minimum rule is used for machine allocation.
[0159] It should be noted that the main process of establishing the main task transfer probability model is as follows: it is mainly divided into two steps, step 1: in the evolutionary process, calculate the evolutionary degree of the first two generations in history; step 2: calculate the possibility of knowledge transfer in the main task population.
[0160] Step 1: During the evolution process, assume that G represents the current generation and t represents the tth task. The optimal objective function value of the main task in the current generation and the previous two generations is expressed as and Compute the difference in the overall optimization process based on the recorded function:
[0161]
[0162]
[0163] In the formula, |.| represents an absolute value. In the process of calculating the optimal value, represents the average Euclidean distance between all individuals in the target space and the origin in the current evolutionary generation, represents the average Euclidean distance between all individuals in the target space and the origin in the previous evolutionary generation, represents the average Euclidean distance between all individuals in the target space and the origin in the first two evolutionary generations.
[0164] Step 2: Calculate the possibility of knowledge transfer in the main task population;
[0165]
[0166] represents the possibility of knowledge transfer in the molecular representation, represents the difference between the best function value in the current generation and the previous generation, Represents the difference between the best function value in the current generation and the previous generation, the larger shows that self-evolution is shrinking, on the contrary, the smaller This indicates that the search is converging in a promising direction.
[0167] The specific operation of individual transfer between multiple tasks is divided into two steps:
[0168] Step 1: Calculate the possibility of knowledge transfer in the main task population;
[0169] Step 2: Determine whether to transfer knowledge individuals. Select individuals from the auxiliary population and migrate them to the main task optimization population as offspring individuals. In the main task optimization population, parent individuals are selected for crossover mutation to produce offspring.
[0170] At the same time, individuals in the auxiliary task are migrated to the main task optimization population, mainly using the similarity calculation of the artifacts, as follows:
[0171]
[0172] In the formula, I i N represents the total number of positions assigned to the same position in two process sequences. i Represents the total number of processes for job i.
[0173] The specific process of transferring individual similarities between tasks is as follows:
[0174] Step 1: Auxiliary task optimization population (CP at ) and the main task optimization population (CP mt ), select two artifact chromosomes in the two types of optimization populations, and calculate the S(i) value of each artifact;
[0175] Step 2: If all values of S(i) are equal to their respective workpiece operation numbers, then randomly identify operation o ij and move it to C i , and shift the affected operations to the right until they reach o ij location;
[0176] On the contrary, if all values of S(i) are equal to 0, half of the assignments are randomly assigned to the subset st, and CP at From st to CP mt The operation of the same location of the job, and the CP that does not belong to the subset mt The operations in are appended to C from left to right i The rest of the locations within;
[0177] Otherwise, based on the decreasing value of S(i), half of the tasks are assigned to the subset st, and CP at From st to CP mt The operation of the same location of the job, and the CP that does not belong to the subset mt The operations in, from left to right, are appended to C i The rest of the location.
[0178] For a machine chain, the process of performing crossover to generate the child machine vectors is as follows:
[0179] Select two random numbers R1 and R2, 1<R1<R2<O(total number of operations), R1 and R2 represent the starting position and ending position of the set of elements to be exchanged, M at With M mtare the machine encoding chromosomes in the auxiliary task population and the main task population respectively. at The elements in R1 are copied from the corresponding positions in R2, and M mt All other elements of M are appended to their respective positions, checking mt The feasibility of the tasks in it and the corresponding operations of the children in it.
[0180] In order to facilitate understanding of the above technical solutions of the present invention, the working principle or operation mode of the present invention in the actual process is described in detail below.
[0181] Based on the classic flexible job shop scheduling example, a job shop scheduling test example with dynamic order changes is designed.
[0182] Reasonable parameter configuration plays a vital role in the performance of the algorithm. The parameter combination selected by this algorithm is P = {60, 80, 100}, Gen = {80, 100, 120}, P c ={0.6,0.7,0.8}, P m ={0.2, 0.3, 0.4}, the parameter orthogonal array L9 (34) is used for calibration experiments, and each parameter configuration is independently run for 20 rounds in the DMK01, DMK08 and DMK14 test cases. According to the comprehensive comparative analysis of the experimental results, the optimal combination of parameters is P = 100, G en =200, P c =0.7, P m =0.4.
[0183] In order to evaluate the effectiveness and superiority of multi-objective optimization algorithms in solving energy scheduling problems, this embodiment uses two indicators, namely, inverted generation distance (IGD) and hypervolume (HV), to evaluate all optimization algorithms.
[0184] In this embodiment, the non-dominated solution obtained by each algorithm running independently is used as the true PF. HV is usually used for evaluation by multi-objective optimization algorithms, and an increase in the HV value indicates an improvement in algorithm performance. IGD is used to evaluate the convergence performance of the algorithm. The lower the IGD value, the better the performance of the algorithm.
[0185] Specifically, IGD is a comprehensive performance evaluation indicator. The smaller the IGD value, the shorter the distance between the optimal solution set obtained by the algorithm and the true Pareto frontier, that is, the better the overall performance of the algorithm. The IGD calculation formula is as follows:
[0186]
[0187] Where P *represents the uniformly distributed point set on the true Pareto frontier, P represents the approximate solution set of the true Pareto frontier obtained by the algorithm, d(v, P) represents the shortest Euclidean distance from point v to all points in P, |p * | represents the set P * The number of individuals in .
[0188] HV mainly evaluates the convergence performance of the population. The larger the HV value, the closer the non-dominated solution in the target space is to the true frontier, and the better the convergence of the algorithm. The calculation formula of HV is as follows:
[0189]
[0190] Where v(x, P) represents the hypervolume of the space formed between the solution x in the non-dominated solution set X and the reference point P corresponding to the true Pareto frontier.
[0191] In ATCOA, the crossover method exchanges information within the population to guide the global evolution of the algorithm, which can enable the population to conduct a large-scale search. Therefore, this embodiment conducts an experimental analysis on the regional method of crossover mutation.
[0192] Method 1: Search only in intersection area 1 in both the main task and the auxiliary task.
[0193] Method 2: In the main task and the auxiliary task, intersection area 2 and intersection area 2 are merged, and intersection area 3 and intersection area 3 are crossed.
[0194] Method three: Search only in intersection area 3 in both the main task and the auxiliary task.
[0195] Method 4: It is also the method adopted in the ATCOA algorithm. In ATCOA, the method of exchanging information between intersection area 1 and intersection area 1, and between intersection area 2 and intersection area 2 is applied in the main task search process, and the method of exchanging information between intersection area 3 and intersection area 3 is applied in the auxiliary task search process.
[0196] Each method was run 20 times independently on the test case. Table 1 shows the experimental results of various methods on the IGD indicator, and the best numerical results are marked in bold.
[0197] Table 1: Experimental results of the crossover method on the IGD index
[0198]
[0199]
[0200] It can be seen from Table 1 that for different crossover regions, using different crossover methods will have different effects on the algorithm. Method 1 obtains the optimal value on 5 test cases, method 2 obtains the optimal value on 2 test cases, and method 3 obtains the optimal value on 1 test case.
[0201] The ATCOA algorithm obtained the optimal value on 7 test cases. By comparing the methods of crossing in different areas, it was found that crossing in each area would make the population crossover more uniform and the search range larger. Therefore, the ATCOA algorithm obtained more excellent experimental results. In method 1, area 1 is at the front end of order changes, which can make the scheduling plan compact in the early stage and shorten the completion time of subsequent processes.
[0202] In order to demonstrate the superiority of the proposed ATCOA algorithm in solving the flexible job shop scheduling problem with order uncertainty, an experimental comparison was conducted between ATCOA and other methods currently used to solve FJSP, namely, the discrete jaya algorithm (Djaya) for solving the new job insertion scheduling problem in FJSP, an improved memetic algorithm (MMA) for order checking, an improved memetic algorithm (RMA) for solving order cancellation, and a fast and efficient multi-objective genetic algorithm (Non-dominated sorting genetic algorithm2, NSGA2) for solving flexible job shop scheduling.
[0203] The parameters of all compared algorithms are the same as the initial settings. All algorithms are independently run for 20 rounds on each test case. The static experimental results of the compared algorithms are shown in Tables 2 and 3. The best numerical results of the algorithms are marked in bold.
[0204] Table 2: Experimental results of the comparison algorithms on the IGD index
[0205]
[0206]
[0207] In Table 2, Djaya algorithm obtained 3 best results on the DMK test case, RMA obtained 1 best value, NSGA2 obtained 1 best value, and ATCOA obtained 10 best results.
[0208] Through the comparison of experimental results, ATCOA obtained the best experimental results in most test cases. When faced with the situation of inserting new orders into the original scheduling plan, the Djaya algorithm generates new solutions in the elite population, the inferior population and the current population, expanding the search range for the generation of new solutions. The Djaya algorithm designed 5 neighborhood search strategies, and further searched on the initial scheduling plan, which may also be the reason why the Djaya algorithm obtained 3 optimal values in the DMK test case.
[0209] In RMA, the classic crossover mutation method is used to generate offspring individuals, and three neighborhood search strategies are used to improve the quality of the solution. The experimental results show that the neighborhood search strategy can improve the performance of the algorithm.
[0210] Table 3: Experimental results of the comparison algorithms on HV indicators
[0211]
[0212]
[0213] In Table 3, the Djaya algorithm obtained 4 optimal values, the NSGA2 algorithm obtained 2 optimal values, and the ATCOA algorithm obtained 9 optimal values. From the experimental results, it can be seen that the performance of each algorithm on the HV index is consistent with that on other indicators, which is enough to illustrate the superiority of the proposed ATCOA algorithm. MMA and RMA can also obtain non-dominated frontiers, but compared with other algorithms, the performance is not very outstanding. Analysis of the superior performance of the ATCOA algorithm may be attributed to multiple aspects, including the multi-task initialization method, the multi-region crossover method, and the strategy of individual migration of auxiliary tasks to the main task, which enables the ATCOA algorithm to obtain the best value on most test cases.
[0214] In order to compare the superiority of the ATCOA algorithm, Figure 6 and Fig.13 The IGD and HV indicator values of all algorithms run in each round of test cases are shown. The overall analysis of the performance of each algorithm shows that Figure 6-Figure 9 In the test, the fluctuation range of the IGD index value of the ATCOA algorithm is small, and the stability of each round of operation is high. The fluctuation range of the NSGA2 algorithm on the DMK05 test case is relatively stable, and the Djaya algorithm is relatively stable in the later operation of the DMK10 test case.
[0215] exist Figures 10 to 13In the optimization process, the ATCOA algorithm has a relatively small fluctuation range. In each independent run, the fluctuation of each algorithm is relatively reasonable, which is consistent with the heuristic nature of random probability search. In the overall comparison of all algorithms, the proposed ATCOA algorithm has achieved good performance indicators in most test cases.
[0216] Therefore, the present invention proposes the ATCOA algorithm to optimize the efficiency and stability of the completion time of the rescheduling scheme for the problem of flexible job shop scheduling with uncertain orders. In the ATCOA method, a real-time order detection framework is established for the situation of order changes. Different rescheduling strategies are adopted to deal with order changes. At the same time, multiple optimization objectives in the workshop scheduling process are considered and converted into multiple optimization tasks, and a multi-task optimization mathematical model is constructed. The present invention designs multi-task population initialization, multi-region crossover method, and multi-task migration optimization strategy for the auxiliary task optimization algorithm. Based on the constructed dynamic test example, the ATCOA algorithm is comprehensively analyzed and compared. The experimental results show that the ATCOA algorithm can effectively search for the optimal solution of the rescheduling scheme, breaking the situation where the traditional optimization algorithm falls into the local optimum, and also achieves the best performance of the comparison algorithm.
[0217] To sum up, with the help of the above-mentioned technical scheme of the present invention, the present invention takes into account the dynamic nature of order changes in the actual production of workshop scheduling, constructs a real-time detection framework with uncertain dynamic order changes, and can promptly respond to dynamic events of order changes occurring in the actual manufacturing process. Compared with the prior art, the present invention is more suitable for handling scheduling problems in intelligent manufacturing workshops in actual production.
[0218] In the face of different types of real-time changes in orders, the present invention designs a dynamic response mechanism to generate a rescheduling plan. In the face of different types of order changes, different dynamic responses are used to generate rescheduling plans, which accelerates the response time for processing dynamic events and improves the feasibility of the rescheduling plan. At the same time, based on the diversity characteristics of the collaborative optimization of multiple tasks in workshop scheduling, the auxiliary task collaborative optimization algorithm is used to solve the solution of the dynamic scheduling plan, which improves the algorithm's solution ability in terms of convergence and diversity, so that a good optimization scheduling plan can be obtained in the optimization process of multiple tasks.
[0219] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm, characterized in that: The dynamic flexible job shop scheduling method comprises the following steps: S1. Build a real-time order change detection framework for dynamic flexible job shop scheduling, and design a rescheduling plan based on the detection results and the real-time processing process of the workshop; S2. Generate a target optimization function based on the rescheduling plan and the workshop optimization task objectives, and build a task optimization mathematical model based on the target optimization function and the workshop conditions; S3. Based on the auxiliary tasks in the dynamic flexible job workshop, the collaborative optimization algorithm is combined with the task optimization mathematical model to optimize the rescheduling plan to obtain the final rescheduling plan in the dynamic flexible job workshop.
2. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 1 is characterized in that: The method of constructing a real-time order change detection framework for the order change in the dynamic flexible job shop scheduling process and designing a rescheduling plan based on the detection results and the real-time processing process of the shop includes the following steps: S11. Execute the pre-scheduling plan at the beginning of the dynamic flexible job shop scheduling, and build a real-time detection framework for order changes during the execution process, and use the detection framework to monitor the execution process of the pre-scheduling plan; S12, judging whether the order has changed dynamically based on the monitoring result, if the order has not changed dynamically, continuing to execute the pre-scheduling plan, if the order has changed dynamically, executing step S13; S13. Determine the corresponding order change type based on the dynamic change status of the order, and select a rescheduling strategy to generate a rescheduling plan according to the order change type; S14. Execute the rescheduling plan, use the real-time order change detection framework to monitor the order status after the rescheduling plan is executed, and determine whether the order has changed dynamically based on the monitoring results.
3. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 2 is characterized in that: The method of judging the corresponding order change type based on the dynamic change status of the order, and selecting the rescheduling strategy to generate the rescheduling plan according to the order change type includes the following steps: S131, acquiring a change type of a corresponding order according to a dynamic change status of the order, wherein the change type includes order insertion, order cancellation, and order insertion and cancellation coexisting; S132. When the order change type is order cancellation, cancel the time node of the corresponding order and reschedule all unprocessed workpieces; S133. When the order change type is order insertion, reschedule the unfinished orders based on the pre-scheduling plan, select a machine for the new order and determine the machine startup time; S134. When the order change type is that order insertion and cancellation exist at the same time, the order scheduling operation is performed based on the workpiece floating time insertion rescheduling strategy.
4. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 3 is characterized in that: When the order change type is that order insertion and cancellation exist at the same time, the order scheduling operation based on the workpiece floating time insertion rescheduling strategy includes the following steps: S1341. When the order change type is order insertion and cancellation coexisting, determine the floating time of the machine corresponding to the new order according to the pre-scheduling plan, where the floating time includes the available idle time, total idle time and gap idle time of the corresponding machine; S1342. If the idle time of the new order is less than or equal to the available idle time, the new order is directly assigned to the corresponding machine position; S1343. If the idle time of the new order is greater than the available idle time and less than or equal to the sum of the available idle time and the gap idle time, the new order is assigned to the corresponding machine and the preset process of the corresponding machine is shifted right; S1344. If the idle time of the new order is greater than the sum of the available idle time and the gap idle time, and less than or equal to the sum of the available idle time and the total idle time, the new order is assigned to the corresponding machine, and the positions of the corresponding workpieces and machines are adjusted; S1345. If the idle time of the new order does not meet all the idle time intervals of the corresponding machines, the standby machine is enabled and the order operation is performed in the order of the new order. If the idle time of the new order does not meet the idle time intervals on all the machines, the process of the new order is placed at the end of the pre-scheduling plan. S1346. Repeat steps S1342 to S1345 to arrange all new orders. If all new orders are arranged, the order arrangement operation is performed. If there are new orders that have not been arranged, repeat steps S1342 to S1345.
5. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 4 is characterized in that: The calculation formula of the corresponding machine gap idle time is: In the formula, Indicates the gap idle time, Indicates the current process O i,j The end time and next process O i,j+1 The difference between the start times, Indicates the current process O i,j The completion time of the current process O i,j Next process O on the machine i,j+1 The difference between the start times, S i,j+1 Indicates the next process O i,j+1 The time to start processing, C ij Indicates the current process O i,j End processing time, r i represents the total number of processes for job i, j represents the process index, C ect Indicates the current process O i,j Estimated completion time, Ci,j represents the current process O i,j The end processing time, m represents the total number of machines in the dynamic flexible job shop, k represents the machine index, and h k represents the number of processes that need to be arranged on the machine, P represents the index of the process position on the machine, S k,P+1 represents the start time of the process at position P+1 on machine k, C k,P represents the end time of the process at position P on machine k, Y i,j,k,P Indicates process o ij Processing time at position P of machine k, C k,hk represents position h on machine k k The process completion time, Y i,j,k,hk Indicates process o ij At position h of machine k k Processing time.
6. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 1 is characterized in that: The generating of the target optimization function based on the rescheduling plan and the workshop optimization task target, and constructing the task optimization mathematical model according to the target optimization function and the workshop status include the following steps: S21, obtaining the completion time, machine load information and workpiece flow time after the order is dynamically changed according to the rescheduling plan, and obtaining the completion time, machine load information and workpiece flow time of the pre-scheduling plan; S22, combining the completion time, machine load information and overall workpiece flow time with the workshop optimization task target to generate a target optimization function; S23. Generate constraint conditions according to the target optimization function, and construct a task optimization mathematical model based on the target optimization function, the constraint conditions and the workshop conditions.
7. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 6 is characterized in that: The process of combining the completion time, machine load information, and overall workpiece flow time with the workshop optimization task objectives to generate a target optimization function includes the following steps: S221, generating a time efficiency objective function and a machine robustness objective function according to the completion times corresponding to the rescheduling plan and the pre-scheduling plan; S222, generating a maximum machine load deviation objective function according to the machine load information corresponding to the rescheduling scheme and the prescheduling scheme; S223, generating an overall workpiece flow time deviation objective function using the overall workpiece flow time corresponding to the rescheduling scheme and the prescheduling scheme; S224, generating a dual-objective optimization function based on the time efficiency objective function, the machine robustness objective function, the maximum machine load deviation objective function and the overall workpiece flow time deviation objective function.
8. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 7 is characterized in that: The expression of the dual objective optimization function is: Where f1 represents the time efficiency objective function, f2 represents the machine robustness objective function, f3 represents the maximum machine load deviation objective function, f4 represents the overall workpiece flow time deviation objective function, and x represents the decision variable.
9. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 2 is characterized in that: The method of combining the collaborative optimization algorithm with the task optimization mathematical model to optimize the rescheduling scheme based on the auxiliary tasks in the dynamic flexible job shop to obtain the final rescheduling scheme in the dynamic flexible job shop includes the following steps: S31, generating an initial population of task optimization based on the pre-scheduling plan, and dividing the tasks into main tasks and auxiliary tasks according to the job requirements of the dynamic flexible job shop; S32. Establish a main task optimization probability model based on the collaborative optimization algorithm, and use the main task optimization probability model to determine whether there is cross-task knowledge transfer; S33. If there is cross-task knowledge transfer, individuals in the auxiliary task are transferred to the main task based on the task knowledge transfer strategy to perform knowledge fusion operation; S34. The dynamic change results of the orders after the execution of the rescheduling plan are combined with the process floating time and the rescheduling decoding strategy inserted according to the task optimization mathematical model, and the rescheduling plan is decoded to obtain the final rescheduling plan of the dynamic flexible job shop.
10. The dynamic flexible job shop scheduling method based on auxiliary task collaborative optimization algorithm according to claim 9 is characterized in that: The task optimization initial population includes a main population and an auxiliary population; The main population is the process with the largest number of remaining processes and the process that follows the rule of allocating the minimum workload to the machine; The auxiliary population is a process that follows the global minimum processing time rule.
Citation Information
Cited By
Simplified fluid relaxation method for dynamic multiple flexible job shop scheduling
CN119668223A
Tire vulcanization workshop intelligent flexible scheduling method and system oriented to dynamic disturbance
CN120891806A
Factory production scheduling decision management system combined with multi-factor intelligent analysis
CN121212693A
