Flexible job shop multi-resource scheduling method, system and device and medium
By constructing the MRC-FJSP production scheduling model and iteratively solving it, combining process coding and machine coding double-layer coding and heuristic rules, the problem that auxiliary resource loading and unloading time is not considered is solved, and an efficient solution for multi-resource scheduling in flexible work workshops is realized , improves production efficiency.
Patent Information
- Application Number
- CN202411167780.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-23
- Publication Date
- 2024-12-03
AI Technical Summary
The existing technology fails to effectively consider the loading and unloading time of auxiliary resources, resulting in a significant increase in the complexity and solution difficulty of multi-resource scheduling problems in flexible operation workshops, which cannot meet actual production needs.
Build an MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time, and obtain the optimal machine selection, processing order and auxiliary resource allocation scheme for each process through iterative solution. The process encoding and machine encoding double-layer encoding method are combined with heuristics Rules determine the allocation of auxiliary resources.
It effectively reduces the loading and unloading time of auxiliary resources, improves resource utilization efficiency, reasonably arranges the allocation of machines and auxiliary resources, and improves production efficiency.
Smart Images

Figure CN119067390A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of workshop scheduling, and in particular to a flexible job shop multi-resource scheduling method, system, equipment and medium. Background Art
[0002] Optimal workshop scheduling is one of the key technologies to achieve high efficiency, high flexibility and high reliability in manufacturing enterprises. The research focuses on "how to allocate limited resources to several tasks within a reasonable time to meet or optimize one or more goals", which has important theoretical and application value for actual production. With the continuous improvement of factory manufacturing flexibility, the classic job shop scheduling problem (Job Shop Scheduling Problem, JSP) cannot fully meet the needs of manufacturing systems. The Flexible Job Shop Scheduling Problem (Flexible Job Shop Scheduling Problem, FJSP) is an extension of JSP. Each process of the workpiece has multiple optional machines, which greatly increases the flexibility and scope of application, but also makes the problem more complicated. At present, domestic and foreign scholars have conducted a lot of research and work on FJSP, and it has been applied in various practical problems.
[0003] Traditional FJSP research focuses on process constraints and equipment resource constraints. However, in the actual production process, workers and other processing resources are also constrained. This type of problem is classified as a dual-resource constrained flexible job shop scheduling problem (DRC-FJSP) when considering both machines and an auxiliary processing resource. If more than one processing resource constraint is considered, it is called a multi-resource constrained flexible job shop scheduling problem (MRC-FJSP). The multi-resource scheduling problem of flexible job shops is widely present in various types of manufacturing enterprises. Taking the production process of a mold factory as an example, the special fixtures required for the processing and assembly processes of different types of molds are different. When processing parts of different types of molds, the processing machine must be equipped with the corresponding fixture and workers with corresponding skills can perform processing. In the production process, it takes time to replace workers, clamp and disassemble fixtures, and replace processing tools. In actual production, the loading and unloading time of resources will have a great impact on scheduling operations. Therefore, the research on multi-resource scheduling problem of flexible job workshop considering loading and unloading is closer to production reality than existing research, but the scale of the problem is further expanded and the difficulty of solving it has increased significantly. Summary of the invention
[0004] In view of the deficiency that the prior art does not take the loading and unloading time of auxiliary resources into consideration, which increases the complexity of process machine selection, processing sequence and allocation of auxiliary resources, the present invention proposes a flexible job shop multi-resource scheduling method, system, equipment and medium. By constructing an MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time, the production scheduling model is iteratively solved to obtain the scheduling solution of machine selection, processing sequence and auxiliary resource allocation for each process, thereby solving the problems existing in the prior art.
[0005] A flexible job shop multi-resource scheduling method comprises the following steps:
[0006] Obtain multiple processes, multiple processing machines and processing time of each process of the workpiece to be processed in the workshop;
[0007] Based on the flexible job shop scheduling FJSP, the constraints of auxiliary resources are added, and the MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time is constructed;
[0008] Input multiple processes, multiple processing machines and the processing time of each process of the workpiece to be processed into the production scheduling model for iterative solution to obtain the optimal scheduling plan for each process; the scheduling plan includes the selection of the processing machine for each process, the processing order of each process and the allocation of auxiliary resources;
[0009] Workshop scheduling is carried out according to the optimal scheduling plan for each process, which specifically includes the following steps:
[0010] The double-layer coding method of process coding and machine coding is used to determine the processing machine for each process and the processing sequence of the processes on the machine;
[0011] By adopting heuristic rules to determine the allocation of auxiliary resources during the decoding process, the workshop scheduling is completed; the adoption of heuristic rules to determine the allocation of auxiliary resources specifically includes: determining the start processing time S of each process on the machine ij ; If the machine is idle, and there is a requirement for at least one piece of each auxiliary resource configuration, select the earliest available time among the auxiliary resources Less than S ij If there are multiple auxiliary resource items that can be used, the earliest auxiliary resource that can be used is selected. The largest single piece; if the auxiliary resource configuration does not have a single piece that can be used, then the processing cannot start immediately. ij When the time available for processing is equal to the time available for auxiliary resources, processing begins.
[0012] Furthermore, the MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time is expressed as:
[0013]
[0014] Among them, C max represents the maximum completion time; C ij Indicates process O ij The processing end time; i and i' are both workpiece numbers; I is a workpiece set; j and j' are both workpiece process numbers; j i Represents the total number of processes for job i.
[0015] Furthermore, the constraints of the MRC-FJSP production scheduling model include:
[0016] The process is processed after the previous process on the same workpiece is completed, which is expressed as Where k represents the machine serial number; K ij Indicates process O ij A collection of processing machines that can be selected; S ij Indicates process O ij Start processing time; X ijk represents a 0-1 decision variable. If process O ij If the processing is selected on machine k, it is 1, otherwise it is 0; p ijk Indicates process O ij Processing time on machine k;
[0017] Each process is processed on the machine in the order of given priority, which is represented by S ij +p ijk ≤S i’j’ +L(3-X ijk -X i’j’k -Y iji’j’k ), where L represents a positive number; Y iji’j’k represents a 0-1 decision variable. If process O ij On machine k, it takes precedence over O i’j’ 1 if processing is done, 0 otherwise;
[0018] A machine can only process one process at a time, which is represented by S i’j’ +p i’j’k ≤S ij +L(2+Y iji’j’k -X ijk -X i’j’k );
[0019] The priority order constraint of the workpiece occupying auxiliary resources for processing is expressed as in, represents a 0-1 decision variable. If process Oij Processing the hth l If there is a single piece, it is 1, otherwise it is 0; represents a 0-1 decision variable. If the hth l A single piece is placed in process O ij The priority is greater than O i’j’ If yes, it is 1, otherwise it is 0. l represents the serial number of the auxiliary resource type;
[0020] A single piece of auxiliary resource can only be configured in the processing of one process at a time, which is expressed as
[0021] At the same time, the process can only be completed on one machine, and the process cannot be interrupted, which is expressed as
[0022] The configuration when all processes are completed uses at least one auxiliary resource, which is represented by Where, f represents the number of auxiliary resource types; R l Indicates the lth auxiliary resource set; h l Indicates the sequence number of the lth auxiliary resource;
[0023] During the processing, multiple auxiliary resources of the same type cannot be configured at the same time, which is expressed as
[0024] The start processing time and processing time of the process cannot be negative, which is represented by S ij ≥0,p ijk ≥0.
[0025] Furthermore, the double-layer coding method of process coding and machine coding is used to determine the processing machine for each process and the processing sequence of the processes on the machine, specifically including:
[0026] The process coding vector OS indicates the priority of process processing, the coding value indicates the workpiece number, and the number of times the coding value appears in the coding vector indicates the process number of the corresponding workpiece;
[0027] The machine coding vector MS represents the processing machine selection of the process, and the coding value represents the machine number selected for the corresponding process.
[0028] Furthermore, the starting processing time S of each process on the machine is determined ij , specifically including the following steps:
[0029] Process O ij The completion time of the corresponding workpiece's preceding process is recorded as A j, the earliest available time of processing machine k is recorded as A k , the total time that auxiliary resources can be used during processing is recorded as A R , then process O ij The start processing time S ij The calculation formula is:
[0030] S ij =max{A j ,A k ,A R}
[0031] If process O ij is the first process of workpiece i, that is, j = 1, then A j =0, if it is not the first process, then:
[0032] A j =C ij-1
[0033] If process O ij is the first process on machine k, then A k = 0, otherwise, A k Depends on the machine k and process O ij Processing process O i’j’ The completion time is:
[0034] A k =C i’j’
[0035] The A R It is expressed as:
[0036]
[0037] in, Indicates the hth auxiliary resource of the lth category l The earliest available time of a single piece, if process O ij This is the first process to configure the auxiliary resource for processing. otherwise The previous process for configuring this auxiliary resource The processing completion time is:
[0038]
[0039] Among them, i * Indicates the workpiece number of the previous process, j * Indicates the sequence number of the previous process.
[0040] Furthermore, the multiple processes of the workpiece to be processed, the multiple processing machines and the processing time of each process are input into the production scheduling model for iterative solution, the optimal solution of the objective function is selected as the elite solution, and the local search strategy is performed on the elite solution to obtain the optimal processing sequence of the process on the machine, which specifically includes the following steps:
[0041] Generate an initialization population; the population represents a set of production scheduling solutions;
[0042] Calculate the fitness values of the individuals in the population; directly copy the top 20% of the individuals with the best fitness in the population to the sub-population, and adopt the tournament selection strategy for the remaining individuals in the population. Each time, a number of individuals are drawn from the original population, and the individual with the best fitness is selected and put into the sub-population; repeat the tournament selection operation until the sub-population size reaches the given population size; the individual represents a production scheduling plan;
[0043] Perform crossover and mutation operations on the selected optimal individuals in turn;
[0044] The variable neighborhood search operation based on the critical path is performed on the first 10% individuals with excellent fitness in the population to obtain the optimal processing sequence of the process on the machine.
[0045] Furthermore, the variable neighborhood search operation based on the critical path is performed on the top 10% individuals with excellent fitness in the population to obtain the optimal processing sequence of the process on the machine, which specifically includes the following steps:
[0046] Decode the individuals that need to perform variable neighborhood search to obtain the processing information of the process, machine, and auxiliary resources of the corresponding scheduling plan;
[0047] Retrieve the end time of the last process on each machine. If the end time is equal to the maximum completion time, create a critical path newPath;
[0048] Determine in sequence whether the machine process and the workpiece process immediately preceding the current process are critical processes. If so, move to the next process for further determination. If not, continue the operation. At the same time, determine whether the start time of the current process is 0. If so, stop searching and add newPath to the set of critical paths.
[0049] Retrieve the process whose end time is equal to the start processing time of the current process, and determine whether it is the auxiliary resource predecessor process; if the same auxiliary resource occupancy exists, add the corresponding process to the critical path and determine again whether the machine predecessor process and the workpiece predecessor process of the current process are critical processes.
[0050] The present invention also includes a flexible job shop multi-resource scheduling system, including:
[0051] An acquisition module is used to acquire multiple processes, multiple processing machines and processing time of each process of the workpiece to be processed in the workshop;
[0052] The scheduling model building module is used to add auxiliary resource constraints based on the flexible job shop scheduling FJSP and build the MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time;
[0053] A calculation module is used to input multiple processes, multiple processing machines and processing time of each process of the workpiece to be processed into the production scheduling model for iterative solution to obtain the optimal scheduling plan for each process; the scheduling plan includes the selection of the processing machine for each process, the processing order of each process and the allocation of auxiliary resources;
[0054] The scheduling module is used to schedule the workshop according to the optimal scheduling plan for each process; specifically, it includes:
[0055] A coding unit is used to determine the processing machine for each process and the processing sequence of the processes on the machine by using a double-layer coding method of process coding and machine coding;
[0056] The decoding unit is used to determine the allocation of auxiliary resources by designing heuristic rules in the decoding process; the heuristic rules specifically include: determining the start processing time S of each process on the machine ij ; If the machine is idle, and there is a requirement for at least one piece of each auxiliary resource configuration, select the earliest available time among the auxiliary resources Less than S ij If there are multiple auxiliary resource items that can be used, the earliest auxiliary resource that can be used is selected. The largest single piece; if the auxiliary resource configuration does not have a single piece that can be used, then the processing cannot start immediately. ij When the time available for processing is equal to the time available for auxiliary resources, processing begins.
[0057] The present invention also includes a flexible job shop multi-resource scheduling computer device, including: a memory, a processor and a computer program stored in the memory, and the processor implements the steps of the flexible job shop multi-resource scheduling method when executing the computer program.
[0058] The present invention also includes a readable storage medium, wherein the readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, they are used to execute the steps of the flexible job shop multi-resource scheduling method.
[0059] The present invention provides a flexible job shop multi-resource scheduling method, system, device and medium, which have the following beneficial effects:
[0060] The present invention constructs a production scheduling model of a multi-resource flexible job shop considering loading and unloading by considering auxiliary resource constraints, and iteratively solves the production scheduling model to obtain the optimal machine selection, processing sequence and auxiliary resource configuration scheme for each process; a double-layer coding method of process coding and machine coding is used to decide the two sub-problems of processing machine selection and process sequencing, and the allocation of auxiliary resources is determined in the decoding process by the designed heuristic rules, thereby establishing a mapping relationship from the coding space to the problem solution space; specifically, the earliest available time is selected close to the processing start time S ij The idle time of auxiliary resources can be compressed, the utilization efficiency of auxiliary resources can be improved, and at the same time, the earliest available time of such auxiliary resources can be advanced, so as to minimize the impact of resource occupation on the subsequent processing steps of configuring such resources for processing, and at the same time, reduce the loading and unloading time of auxiliary resources while ensuring the feasibility of the scheduling solution; the present invention can efficiently solve the multi-resource scheduling problem of a flexible job shop considering loading and unloading, reasonably arrange the configuration of machines and auxiliary resources, reduce resource loading and unloading, and improve production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0061] Figure 1 It is a structural diagram of the production scheduling model MRC-FJSP in an embodiment of the present invention;
[0062] Figure 2 This is a schematic diagram of individual coding in an embodiment of the present invention;
[0063] Figure 3 A schematic diagram of a processing order and a machine selection sequence obtained by decoding in an embodiment of the present invention;
[0064] Figure 4 A machine Gantt chart in an embodiment of the present invention;
[0065] Figure 5 A Gantt chart of auxiliary resources in an embodiment of the present invention;
[0066] Figure 6 Schematic diagram of the structure of the neighborhood structure 1 in an embodiment of the present invention;
[0067] Figure 7 This is a schematic diagram of the structure of a neighborhood structure 2 in an embodiment of the present invention;
[0068] Figure 8 This is a schematic diagram of the structure of the neighborhood structure 3 in an embodiment of the present invention;
[0069] Fig. 9 Schematic diagram of the process of a genetic variable neighborhood search algorithm based on a critical path in an embodiment of the present invention. DETAILED DESCRIPTION
[0070] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.
[0071] The present invention proposes a flexible job shop multi-resource scheduling method considering the loading and unloading time of auxiliary resources, and constructs an MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time; MRC-FJSP is an extension based on FJSP. FJSP is described as: there are n workpieces {J1, J2, ..., J n}On m machines {M1,M2,...,M m} for processing. Workpiece J i Contains i The process route and process priority of the workpiece are predetermined. ij The number of machines in the optional processing machine set is greater than or equal to 1, and the corresponding processing time is fixed and known. The goal of solving FJSP is to determine the processing machine selection for each process and the processing order of the processes on each machine so that the selected performance indicators (such as the maximum completion time C max ) is optimal. FJSP follows the following assumptions:
[0072] (1) Each machine can only process one process at a time.
[0073] (2) Each workpiece can only be processed by one machine at a time.
[0074] (3) At time zero, all machines are available and all workpieces can begin processing.
[0075] (4) No interruption is allowed after the process starts.
[0076] (5) The processing priorities of processes belonging to different workpieces are the same.
[0077] MRC-FJSP takes into account the constraints of auxiliary resources on the basis of classic FJSP. Therefore, solving MRC-FJSP requires not only determining the processing machine for each process and the processing order of the processes on the machine, but also determining the allocation of auxiliary resources, such as Figure 1 Based on the FJSP problem description, MRC-FJSP needs to add the following assumptions:
[0078] (1) The processing process requires h kinds of auxiliary resources {R1, R2, ..., R h}, the types of auxiliary resources and the number of each type of auxiliary resources are predetermined.
[0079] (2) Each machine must meet the configuration of a given type of auxiliary resources before it can perform processing.
[0080] (3) Auxiliary resources of the same type have the same priority.
[0081] (4) At time zero, all auxiliary resources are available.
[0082] (5) Each auxiliary resource can only be used on one machine at a time.
[0083] In order to facilitate the description of the problem characteristics of MRC-FJSP, such as Figure 1 As shown, the present invention gives a simple example for illustration. In this example, there are 3 workpieces processed on 3 machines. The processing process requires the configuration of 3 auxiliary resources, and each auxiliary resource has 2 identical units. Table 1 gives the workpiece processing information of the example, including the process of the workpiece, the set of optional processing machines for the process, the processing time of the process on the machine, and the auxiliary resource configuration required for the process processing. Taking the first process of the workpiece J1 in the example as an example, it can be selected to be processed in M1 and M3, and the corresponding processing times are 4 and 3 respectively, and the types of auxiliary resources that need to be configured to complete the processing are 1 and 3.
[0084] Table 1 Example processing information
[0085]
[0086] The scheduling method includes the following steps:
[0087] S1. According to the problem description, the MRC-FJSP mathematical model with the goal of minimizing the maximum completion time is established. The parameter symbols and decision variables used in the model are defined as shown in Table 2:
[0088] Table 2 MRC-FJSP mathematical model parameter symbols and decision variable definitions
[0089]
[0090]
[0091] Decision variables:
[0092]
[0093] Based on the defined variables and symbols, the mathematical model of MRC-FJSP is established as follows:
[0094] Objective function:
[0095]
[0096] Constraints:
[0097]
[0098] Where:
[0099] S ij +p ijk ≤S i’j’ +L(3-X ijk -X i’j’k -Y iji’j’k ) (3)
[0100] Where:
[0101] S i’j’ +p i’j’ ≤S ij +L(2+Y iji’j’k -X ijk -X i’j’k ) (4)
[0102] Where:
[0103]
[0104] Where:
[0105]
[0106] Where:
[0107]
[0108] Where:
[0109]
[0110] Where:
[0111]
[0112] Where:
[0113]
[0114] Where:
[0115] S ij ≥0,p ijk ≥0 (11)
[0116] Where:
[0117] Formula (1) gives the definition of the optimization objective of minimizing the maximum completion time; Formula (2) ensures the process constraint requirements of the workpiece, that is, the process can only be processed after the previous process on the same workpiece is completed; Formula (3) ensures that each process is processed on the machine in the order of given priority; Formula (4) ensures that a machine can only process one process at a time; Formula (5) represents the priority order constraint of the workpiece occupying auxiliary resources for processing; Formula (6) ensures that a single auxiliary resource can only be configured in the processing process of one process at a time; Formulas (7) and (8) ensure that a process can only be completed on one machine at the same time, and the process cannot be interrupted; Formula (9) indicates that all processes need to be completed with at least one auxiliary resource; Formula (10) indicates that multiple auxiliary resources of the same type cannot be configured at the same time during the processing; Formula (11) is a non-negative constraint, that is, the start processing time and processing time of the process cannot be negative.
[0118] S2. Carry out workshop scheduling according to the scheduling plan of machine selection, processing sequence and auxiliary resource allocation for each process.
[0119] S2.1. In the study of metaheuristic algorithms for solving combinatorial optimization problems, choosing a suitable encoding method can reduce the invalid search of the algorithm in the solution space and improve the efficiency of the solution. MRC-FJSP can be decomposed into three sub-problems: the selection of process processing machines, the processing order of processes on the machines, and the configuration of auxiliary resources. As a special MRC-FJSP problem, the related research of DRC-FJSP often uses a hierarchical encoding method to encode the three sub-problems separately. For DRC-FJSP, since only one auxiliary resource is considered in addition to the processing machine, three layers of encoding can represent a complete scheduling solution for the three sub-problems, and the size of the solution space is within the scope of the algorithm's solution capability. In MRC-FJSP, multiple auxiliary resources are considered. If the auxiliary resource configuration sub-problem is encoded separately, each additional auxiliary resource needs to design an additional layer of independent encoding decision resource allocation. When considering a large number of auxiliary resources, encoding the auxiliary resources separately will make the solution space too large, and the algorithm cannot obtain a satisfactory scheduling solution within an acceptable time. Therefore, the present invention adopts the double-layer coding method of process coding and machine coding commonly used in FJSP research to decide the two sub-problems of processing machine selection and process sorting. The allocation of auxiliary resources is determined during the decoding process through the designed heuristic rules, thereby establishing a mapping relationship from the coding space to the problem solution space.
[0120] by Figure 2 The following is an example to introduce the encoding method used. The process encoding vector OS represents the priority of the process. In this example, the OS process encoding vector {1,3,2,1,2,2,3,1} represents the order of process arrangement and processing. 11 ,O 31,O 21 ,O 12 ,O 22 ,O 23 ,O 32 ,O 13 The machine encoding vector MS represents the processing machine selection of the process. In the example, the MS machine encoding vector {2,1,2,4,1,3,4,2} represents the process {0 11 ,O 12 ,O 13 ,O 21 ,O 22 ,O 23 ,O 31 ,O 32}The selected processing machines are {M2,M1,M2,M4,M1,M3,M4,M2}.
[0121] S2.2. Heuristic rules for decoding and auxiliary resource selection:
[0122] The decoding operation converts the coding vector into a specific scheduling scheme and calculates the objective function value corresponding to the scheduling scheme. The decoding method needs to be designed in combination with the coding method and the problem characteristics. Since the coding method only considers machine selection and process sorting, and does not consider the auxiliary resource configuration problem, the present invention designs a heuristic rule for auxiliary resource selection in the decoding process. The auxiliary resource selection rule will be introduced in combination with the decoding process.
[0123] According to the problem description of MRC-FJSP, the process can be processed at the current moment only if three prerequisites are met at the same time:
[0124] The preceding process of the workpiece corresponding to the process is completed. If the process is the first process of the workpiece, there is no preceding process constraint.
[0125] The corresponding processing machine is currently idle.
[0126] At present, at least one of all the auxiliary resources required for processing is available.
[0127] The core of the decoding operation is to determine the start time of each process on the machine. ij The completion time of the corresponding workpiece's preceding process is recorded as A j , the earliest available time of processing machine k is recorded as A k , the time when all auxiliary resources required for processing are available is recorded as A R According to the above process constraints, machine constraints and auxiliary resource constraints, process O ij The start processing time S ij It can be calculated by formula (12):
[0128] Sij =max{A j ,A k ,A R} (12)
[0129] If process O ij is the first process of workpiece i, that is, j = 1, then A j =0, if it is not the first process, then:
[0130] A j =C ij-1 (13)
[0131] If process O ij is the first process on machine k, with A k = 0, otherwise, A k Depends on the process O on machine k ij Processing process O i’j’ The completion time is:
[0132] A k =C i’j’ (14)
[0133] A R It indicates the time when all auxiliary resources required for processing can be configured, which can be obtained by formula (15):
[0134]
[0135] In the formula Indicates the hth auxiliary resource of the lth category l The earliest available time of a single piece, if process O ij This is the first process to configure the auxiliary resource for processing. otherwise The previous process for configuring this auxiliary resource The processing completion time is:
[0136]
[0137] Using the MRC-FJSP problem example given in Table 1, which includes 4 artifacts, 3 machines, and 3 types of auxiliary resources, the decoding method and auxiliary resource selection rules are introduced in detail. Figure 3 The code representing a scheduling solution traverses the process code OS from left to right, and determines the processing machine assigned to each process through the machine code MS. After determining the processing order and machine selection sequence, it is also necessary to determine the sub-problem of auxiliary resource selection to obtain a complete scheduling solution. The configuration of auxiliary resources is determined by heuristic rules. The specific selection method is: if the machine is idle, each auxiliary resource that needs to be configured has at least one unit available, and the available time is selected Less than S ij If there are multiple available auxiliary resource items of the same type, select the available time If there is no available single piece of auxiliary resource to be configured, processing cannot start immediately and needs to wait for the auxiliary resource to be available, that is, the processing start time S ij =A R , all auxiliary resources that need to be configured have at least one available, and the selection method of auxiliary resource pieces is the same as the above case. The reason for applying this scheduling rule to determine the selection of auxiliary resource pieces is that there is no difference between auxiliary resource pieces of the same type, and the earliest available time is selected close to the processing start time S ij It can compress the idle time of auxiliary resources, improve the utilization efficiency of auxiliary resources, and at the same time advance the earliest available time of such auxiliary resources, so as to minimize the impact of resource occupation on the subsequent processing steps of such resources. Figure 3 Taking the encoding in as an example, the double-layer encoding of process sequence and machine selection is decoded to obtain the processing sequence and machine selection corresponding to the process. In the decoding process, the configuration sub-problem of auxiliary resources is dynamically decided to obtain a complete scheduling solution. The Gantt chart of machines and auxiliary resources corresponding to the scheduling solution is shown in Figure 4 and Figure 5 As shown, process O 12 The processing machine is M1, and the auxiliary resources to be configured are R2 and R3. 12 The preceding process O 21 The completion time is 6, and the workpiece J1 is on O 12 Pre-process O 11 The completion time is 3. At time 6, the No. 1 unit of auxiliary resource R2 is available, and the No. 1 and No. 2 units of R3 are both available, which meets the prerequisite for starting processing. Since R2 has only one auxiliary resource unit available, there is no need to select it. Auxiliary resource R2 selects the idle No. 1 unit. However, auxiliary resource R3 has two units, No. 1 and No. 2, to choose from. The earliest available time of No. 1 of R3 is 3, and the earliest available time of No. 2 is 6. According to the heuristic rule for auxiliary resource selection, No. 2, which has a later available time, is selected and configured in process O. 12 In the process of R3, the earliest available time is the moment 3 of unit 1. If the decision is not made by the auxiliary resource selection rule, and unit 1 of R3 is randomly selected, the earliest available time of R3 is the moment 6 of unit 2, which is later than the scheduling scheme selected by the scheduling rule. The earliest available time of auxiliary resource R3 is delayed, so selecting auxiliary resources by rules is better than random selection. 41 The processing machine is M2. At time 8, machine M2 is idle. 41The resources R2 and R3 required for processing do not have available units, so they need to wait for process O. 12 At time 9, after the processing is completed and pieces 1 of R2 and 2 of R3 are available, process O 41 Only then can processing be performed on M2.
[0138] S3, such as Fig. 9 As shown, population initialization method: use random initialization method to generate a series of initial solutions.
[0139] Selection, crossover and mutation operators: The selection operation uses the elite retention strategy and the tournament selection strategy. First, the top 20% of individuals with excellent fitness in the population are directly copied to the sub-population, and the remaining individuals in the population are selected using the tournament selection strategy. Each time, several individuals are drawn from the original population, and the individual with the best fitness is selected and placed in the sub-population. Repeat the tournament selection operation until the sub-population size reaches the given population size.
[0140] The crossover adopts the POX crossover method; the mutation adopts the single-point mutation method.
[0141] S3.1, variable neighborhood search:
[0142] A critical path-based variable neighborhood search strategy was designed, and a critical path-based variable neighborhood search operation was performed on the top 10% individuals with excellent fitness in the population.
[0143] Critical path information acquisition:
[0144] Step 1: Decode the individual to obtain the process, machine, and auxiliary resource processing information of the corresponding scheduling plan;
[0145] Step 2: Retrieve the end time of the last process on each machine. If the end time is equal to makespa n, create a critical path newPath;
[0146] Step 3: Determine in turn whether the machine process and the workpiece process immediately preceding the current process are critical processes. If so, move to the next process and repeat the operation of Step 3. If not, continue the operation of Step 4. At the same time, determine whether the start time of the current process is 0. If it is 0, stop searching and add newPath to the criticalPath set.
[0147] Step 4: Retrieve the process whose end time is equal to the start time of the current process, determine whether it is the previous process of auxiliary resources, that is, whether there is the same auxiliary resource occupancy, add the corresponding process to the critical path and repeat Step 3.
[0148] Neighborhood structure 1: Neighborhood structure 1 is a commonly used neighborhood structure for replacing machines, such as Figure 6 As shown; the specific operation is to randomly select a key process. If there are multiple optional processing machines for this process, the process is moved to a random machine other than the original machine in the optional machine set for processing; otherwise, the operation process is reselected.
[0149] Neighborhood structure 2: Neighborhood structure 2 is completed by two processes of exchanging the head and tail of the key block, such as Figure 7 As shown; the specific operation is to randomly select a key block with a process number ≥ 2. If it is the first key block, only the two key processes at the end of the block are exchanged; if it is the last key block, only the first process of the block is exchanged; for the key block in the middle, the processes at the beginning and end of the block are exchanged.
[0150] Neighborhood structure 3 is designed based on the idea of reducing the loading and unloading time of auxiliary resources. Different from the general FJSP, due to the existence of auxiliary resources and loading and unloading time, exchanging the processes in the key block may also reduce the maximum completion time. A key process is selected and inserted into other positions in the key block, such as Figure 8 shown.
[0151] Based on the same inventive concept, the present invention also proposes a flexible job shop multi-resource scheduling system, comprising:
[0152] The acquisition module is used to obtain multiple processes, multiple processing machines and the processing time of each process of the workpiece to be processed in the workshop.
[0153] The scheduling model building module is used to add constraints on auxiliary resources based on the flexible job shop scheduling FJSP, and to build a MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time.
[0154] The calculation module is used to input multiple processes, multiple processing machines and the processing time of each process of the workpiece to be processed into the production scheduling model for iterative solution to obtain the optimal scheduling plan for each process; the scheduling plan includes the selection of processing machines for each process, the processing order of each process and the allocation of auxiliary resources.
[0155] The scheduling module is used to schedule the workshop according to the optimal scheduling plan for each process; specifically, it includes:
[0156] A coding unit is used to determine the processing machine for each process and the processing sequence of the processes on the machine by using a double-layer coding method of process coding and machine coding;
[0157] A decoding unit is used to determine the allocation of auxiliary resources by adopting heuristic rules in the decoding process; the heuristic rules specifically include: determining the start processing time S of each process on the machine ij ; If the machine is idle, and there is a requirement for at least one piece of each auxiliary resource configuration, select the earliest available time among the auxiliary resources Less than S ij If there are multiple auxiliary resource items that can be used, the earliest auxiliary resource that can be used is selected. The largest single piece; if the auxiliary resource configuration does not have a single piece that can be used, then the processing cannot start immediately. ij When the time available for processing is equal to the time available for auxiliary resources, processing begins.
[0158] Based on the same inventive concept, the present invention also proposes a flexible job shop multi-resource scheduling computer device, including: a memory, a processor and a computer program stored in the memory, and the processor implements the steps of the flexible job shop multi-resource scheduling method when executing the computer program.
[0159] Based on the same inventive concept, the present invention also proposes a readable storage medium, which stores a computer program. The computer program includes program instructions. When the program instructions are executed by a processor, they are used to execute the steps of the flexible job shop multi-resource scheduling method.
[0160] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A flexible job shop multi-resource scheduling method, characterized in that: The following steps are involved: Obtain multiple processes, multiple processing machines and processing time of each process of the workpiece to be processed in the workshop; Based on the flexible job shop scheduling FJSP, the constraints of auxiliary resources are added, and the MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time is constructed; Input multiple processes, multiple processing machines and the processing time of each process of the workpiece to be processed into the production scheduling model for iterative solution to obtain the optimal scheduling plan for each process; the scheduling plan includes the selection of the processing machine for each process, the processing order of each process and the allocation of auxiliary resources; Workshop scheduling is carried out according to the optimal scheduling plan for each process, which specifically includes the following steps: The double-layer coding method of process coding and machine coding is used to determine the processing machine for each process and the processing sequence of the processes on the machine; By adopting heuristic rules to determine the allocation of auxiliary resources during the decoding process, the workshop scheduling is completed; the adoption of heuristic rules to determine the allocation of auxiliary resources specifically includes: determining the start processing time S of each process on the machine ij ; If the machine is idle, and there is a requirement for at least one piece of each auxiliary resource configuration, select the earliest available time among the auxiliary resources Less than S ij If there are multiple auxiliary resource items that can be used, the earliest auxiliary resource that can be used is selected. The largest single piece; if the auxiliary resource configuration does not have a single piece that can be used, then the processing cannot start immediately. ij When the time available for processing is equal to the time available for auxiliary resources, processing begins.
2. A flexible job shop multi-resource scheduling method according to claim 1, characterized in that: The MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time is expressed as: Among them, C max represents the maximum completion time; C ij Indicates process O ij The processing end time; i and i' are both workpiece numbers; I is a workpiece set; j and j' are both workpiece process numbers; j i Represents the total number of processes for job i.
3. A flexible job shop multi-resource scheduling method according to claim 2, characterized in that: The constraints of the MRC-FJSP production scheduling model include: The process is processed after the previous process on the same workpiece is completed, which is expressed as Where k represents the machine serial number; K ij Indicates process O ij A collection of processing machines that can be selected; S ij Indicates process O ij Start processing time; X ijk represents a 0-1 decision variable. If process O ij If the processing is selected on machine k, it is 1, otherwise it is 0; p ijk Indicates process O ij Processing time on machine k; Each process is processed on the machine in the order of given priority, which is represented by S ij +p ijk ≤S i’j’ +L(3-X ijk -X i’j’k -Y iji’j’k ), where L represents a positive number; Y iji’j’k represents a 0-1 decision variable. If process O ij On machine k, it takes precedence over O i’j’ 1 if processing is done, 0 otherwise; A machine can only process one process at a time, which is represented by S i’j’ +p i’j’k ≤S ij +L(2+Y iji’j’k -X ijk -X i’j’k ); The priority order constraint of the workpiece occupying auxiliary resources for processing is expressed as in, represents a 0-1 decision variable. If process O ij Processing the hth l If there is a single piece, it is 1, otherwise it is 0; represents a 0-1 decision variable. If the hth l A single piece is placed in process O ij The priority is greater than O i’j’ If yes, it is 1, otherwise it is 0. l represents the serial number of the auxiliary resource type; A single piece of auxiliary resource can only be configured in the processing of one process at a time, which is expressed as At the same time, the process can only be completed on one machine, and the process cannot be interrupted, which is expressed as The configuration when all processes are completed uses at least one auxiliary resource, which is represented by Where, f represents the number of auxiliary resource types; R l Indicates the lth auxiliary resource set; h l Indicates the sequence number of the lth auxiliary resource; During the processing, multiple auxiliary resources of the same type cannot be configured at the same time, which is expressed as The start processing time and processing time of the process cannot be negative, which is represented by S ij ≥0,p ijk ≥0.
4. A flexible job shop multi-resource scheduling method according to claim 1, characterized in that: The double-layer coding method of process coding and machine coding is used to determine the processing machine for each process and the processing sequence of the processes on the machine, specifically including: The process coding vector OS indicates the priority of process processing, the coding value indicates the workpiece number, and the number of times the coding value appears in the coding vector indicates the process number of the corresponding workpiece; The machine coding vector MS represents the processing machine selection of the process, and the coding value represents the machine number selected for the corresponding process.
5. A flexible job shop multi-resource scheduling method according to claim 2, characterized in that: The starting processing time S of each process on the machine is determined ij , specifically including the following steps: Process O ij The completion time of the corresponding workpiece's preceding process is recorded as A j , the earliest available time of processing machine k is recorded as A k , the total time that auxiliary resources can be used during processing is recorded as A R , then process O ij The start processing time S ij The calculation formula is: S ij =max{A j ,A k ,A R } If process O ij is the first process of workpiece i, that is, j = 1, then A j =0, if it is not the first process, then: A j =C ij-1 If process O ij is the first process on machine k, then A k = 0, otherwise, A k Depends on the machine k and process O ij Processing process O i’j’ The completion time is: A k =C i’j’ The A R It is expressed as: in, Indicates the hth auxiliary resource of the lth category l The earliest available time of a single piece, if process O ij This is the first process to configure the auxiliary resource for processing. otherwise The previous process for configuring this auxiliary resource The processing completion time is: Among them, i * Indicates the workpiece number of the previous process, j * Indicates the sequence number of the previous process.
6. A flexible job shop multi-resource scheduling method according to claim 1, characterized in that: The method of inputting multiple processes, multiple processing machines and the processing time of each process of the workpiece to be processed into the production scheduling model for iterative solution, selecting the optimal solution of the objective function as the elite solution, and performing a local search strategy on the elite solution to obtain the optimal processing sequence of the process on the machine, specifically includes the following steps: Generate an initialization population; the population represents a set of production scheduling solutions; Calculate the fitness values of the individuals in the population; directly copy the top 20% of the individuals with the best fitness in the population to the sub-population, and adopt the tournament selection strategy for the remaining individuals in the population. Each time, a number of individuals are drawn from the original population, and the individual with the best fitness is selected and put into the sub-population; repeat the tournament selection operation until the sub-population size reaches the given population size; the individual represents a production scheduling plan; Perform crossover and mutation operations on the selected optimal individuals in turn; The variable neighborhood search operation based on the critical path is performed on the first 10% individuals with excellent fitness in the population to obtain the optimal processing sequence of the process on the machine.
7. A flexible job shop multi-resource scheduling method according to claim 6, characterized in that: The method of performing a variable neighborhood search operation based on a critical path on the top 10% individuals with excellent fitness in the population to obtain the optimal processing sequence of the process on the machine specifically includes the following steps: Decode the individuals that need to perform variable neighborhood search to obtain the processing information of the process, machine, and auxiliary resources of the corresponding scheduling plan; Retrieve the end time of the last process on each machine. If the end time is equal to the maximum completion time, create a critical path newPath; Determine in sequence whether the machine process and the workpiece process immediately preceding the current process are critical processes. If so, move to the next process for further determination. If not, continue the operation. At the same time, determine whether the start time of the current process is 0. If so, stop searching and add newPath to the set of critical paths. Retrieve the process whose end time is equal to the start processing time of the current process, and determine whether it is the auxiliary resource predecessor process; if the same auxiliary resource occupancy exists, add the corresponding process to the critical path and determine again whether the machine predecessor process and the workpiece predecessor process of the current process are critical processes.
8. A flexible job shop multi-resource scheduling system, characterized in that: include: An acquisition module is used to acquire multiple processes, multiple processing machines and processing time of each process of the workpiece to be processed in the workshop; The scheduling model building module is used to add auxiliary resource constraints based on the flexible job shop scheduling FJSP and build the MRC-FJSP production scheduling model with the goal of minimizing the maximum completion time; A calculation module is used to input multiple processes, multiple processing machines and processing time of each process of the workpiece to be processed into the production scheduling model for iterative solution to obtain the optimal scheduling plan for each process; the scheduling plan includes the selection of the processing machine for each process, the processing order of each process and the allocation of auxiliary resources; The scheduling module is used to schedule the workshop according to the optimal scheduling plan for each process; specifically, it includes: A coding unit is used to determine the processing machine for each process and the processing sequence of the processes on the machine by using a double-layer coding method of process coding and machine coding; The decoding unit is used to determine the allocation of auxiliary resources by designing heuristic rules in the decoding process; the heuristic rules specifically include: determining the start processing time S of each process on the machine ij ; If the machine is idle, and there is a requirement for at least one piece of each auxiliary resource configuration, select the earliest available time among the auxiliary resources Less than S ij If there are multiple auxiliary resource items that can be used, the earliest auxiliary resource that can be used is selected. The largest single piece; if the auxiliary resource configuration does not have a single piece that can be used, then the processing cannot start immediately. ij When the time available for processing is equal to the time available for auxiliary resources, processing begins.
9. A flexible job shop multi-resource scheduling computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory, wherein when the processor executes the computer program, the steps of the flexible job shop multi-resource scheduling method according to any one of claims 1 to 7 are implemented.
10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by a processor, they are used to execute the steps of the flexible job shop multi-resource scheduling method according to any one of claims 1 to 7.
Citation Information
Cited By
Automatic suspension conveying system for bicycle frames
CN121168732A
Large language model assisted flexible job shop scheduling algorithm
CN122509527A
