A method for solving flexible shop scheduling problem based on improved golden jackal optimization algorithm

Abstract

The present invention provides a method for solving a flexible workshop scheduling problem based on an improved golden jackal optimization algorithm, including: generating an initialization population based on a combination of GLS initialization and RI initialization, and calculating the fitness values ​​of individuals in the initialization population; comparing the fitness values ​​of the individuals, and determining male golden jackals and female golden jackals according to the comparison results; setting the current maximum number of iterations to 0, respectively updating the positions of a machine selection part and a process sorting part, and iterating according to the position update results, and updating the current optimal solution based on the iterative processing results; judging whether the current iteration meets the iteration termination condition, if the iteration termination condition is met, outputting the optimal solution, otherwise executing step 3 for the next iteration. While improving the local search capability of the golden jackal optimization algorithm, the global search capability is improved to ensure that the best solution that meets the global deployment of the workshop is provided for the flexible workshop scheduling problem.

Description

Technical Field

[0001] The present invention relates to the technical field of scheduling control, and in particular to a method for solving flexible workshop scheduling problems based on an improved golden jackal optimization algorithm. Background Art

[0002] As an extension of the classic job shop scheduling problem (JSP), the Flexible Job Shop Scheduling Problem (FJSP) is widely used in manufacturing operations, chemical industry, assembly and other fields. FJSP is more difficult than JSP because it relaxes the restrictions on processing machines and introduces the decision-making content of the job path. Since FJSP is closer to the actual needs of current factory production, it has become a hot topic in the field of production and assembly scheduling problems in recent years.

[0003] The Golden Jackal Optimization Algorithm (GJO) is one of the swarm intelligence optimization algorithms for solving the flexible workshop scheduling problem. Its main feature is that the entire population is divided into two parts: the golden jackal group and the prey group, and the individual positions of the population are updated in the two stages of searching for prey and encircling prey, so that the algorithm takes into account both global search and local search. The algorithm has certain defects. Since the update of the prey position depends on the male jackal, and under certain conditions, the diversity of the golden jackal is limited, the algorithm is easy to fall into the local optimum. Therefore, the present invention proposes a method for solving the flexible workshop scheduling problem based on the improved golden jackal optimization algorithm. Summary of the invention

[0004] The present invention provides a method for solving the flexible workshop scheduling problem based on an improved golden jackal optimization algorithm, so as to solve the above-mentioned problem.

[0005] The present invention provides a method for solving the flexible workshop scheduling problem based on an improved golden jackal optimization algorithm, comprising:

[0006] Step 1: Generate an initialization population based on the combination of GLS initialization and RI initialization, and calculate the fitness values ​​of individuals in the initialization population;

[0007] Step 2: Compare the fitness values ​​of the individuals, and determine the male golden jackal and the female golden jackal based on the comparison results;

[0008] Step 3: Set the current maximum number of iterations to 0, update the positions of the machine selection part and the process sorting part respectively, perform iterative processing according to the position update results, and update the current optimal solution based on the iterative processing results;

[0009] Step 4: Determine whether the current iteration meets the iteration termination condition. If so, output the optimal solution. Otherwise, execute step 3 for the next iteration.

[0010] Preferably, in a method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm, before executing step 1, the method includes:

[0011] Get relevant iteration parameter setting information;

[0012] The relevant iteration parameters include population size, maximum number of iterations, crossover probability factor, mutation probability factor and initialization rule fusion ratio.

[0013] Preferably, in a method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm, before executing step 1, the method includes:

[0014] The golden jackal optimization algorithm is improved based on a new nonlinear convergence factor and the golden sine algorithm.

[0015] Preferably, in a method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm, step 3 specifically includes:

[0016] Calculate the fitness value of the individual after the position is updated according to the decoding strategy, and if the fitness value is greater than the current individual, replace the current individual based on the individual after the position is updated;

[0017] Perform a crossover mutation operation on the replaced individual, and calculate a new solution of the crossover mutation operation according to the decoding strategy. If the new solution is better than the solution corresponding to the current individual, the new solution is accepted as the optimal solution;

[0018] A variable neighborhood search strategy is executed on the optimal solution to generate a new neighborhood solution and evaluate it. If the new neighborhood solution is better than the current optimal solution, the current optimal solution is updated.

[0019] Preferably, in a method for solving the flexible workshop scheduling problem based on the improved golden jackal optimization algorithm, the fitness value of the individual after the position is updated is calculated according to the decoding strategy. If the fitness value is greater than the current individual, the Cauchy-Gauss mutation strategy is used to randomly perturb the optimal individual generated in each iteration in the process of replacing the current individual based on the individual after the position is updated.

[0020] Preferably, in a method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm, step 1 specifically includes:

[0021] Respectively obtain the machine selection fusion ratios of multiple first target selection rules included in the GLS initialization;

[0022] Respectively obtain the process ranking fusion ratios of multiple second target selection rules included in RI initialization;

[0023] According to the machine selection fusion ratio, the first target selection rule is fused to generate a GLS initialization rule, and the machine selection part of the original population is initialized based on the GLS initialization rule to obtain the machine selection part initialization result;

[0024] According to the process sorting fusion ratio, the second target selection rules are fused to generate RI initialization rules, and the process sorting part of the original population is initialized based on the RI initialization rules to obtain the initialization result of the process sorting part;

[0025] Based on the initialization results of the machine selection part and the initialization results of the process sorting part, an initialization population is obtained;

[0026] Based on the genetic algorithm, the fitness values ​​of individuals in the initialized population are calculated.

[0027] Preferably, in a method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm, after executing step 1, the method includes:

[0028] Based on the initial population data, the iteration setting parameters are intelligently modified, including:

[0029] Based on the population size setting information, determine the number of production-adaptive machines for the products to be produced and the types of products to be produced;

[0030] Based on the type of product to be produced, obtain the production process of the product to be produced, determine the production process of the product to be produced, and the sequential relationship between the processes;

[0031] According to the number of production adaptation machines and the sequential association relationship, obtain the historical scheduling processing data corresponding to the machines of similar historical production products;

[0032] Based on historical scheduling processing data, predict the best scheduling strategy for products to be produced;

[0033] According to the optimal scheduling strategy, the optimal solution position is predicted for the current initialized population, and the number of iteration terminations is determined based on the prediction results;

[0034] and comparing the number of iteration terminations with the maximum number of iterations in the relevant iteration parameter setting information to obtain a setting error degree, and when the setting error degree is greater than a preset value, modifying the maximum number of iterations based on the number of iteration terminations;

[0035] Otherwise, the number of iteration terminations remains unchanged.

[0036] Preferably, in a method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm, the optimal scheduling strategy for the product to be produced is predicted based on historical scheduling processing data, including:

[0037] Based on historical scheduling processing data, multiple optimal selection rankings are obtained, and multiple different production scheduling sequences are determined based on the optimal selection rankings;

[0038] Compare multiple different production scheduling sequences to obtain the same scheduling subsequences and different scheduling subsequences;

[0039] Extract scheduling features for different scheduling subsequences respectively, obtain scheduling similar features and compare the features, and schedule similar features;

[0040] Based on the similarity characteristics of the scheduling and the local process characteristics corresponding to the same scheduling subsequence, the sorting order of the initialized population is adjusted to obtain the optimal scheduling strategy.

[0041] Preferably, in a method for solving a flexible shop scheduling problem based on an improved golden jackal optimization algorithm, before executing step 1, the method includes: after a user inputs original population information, obtaining operation and maintenance data of each machine in the original population, and judging whether there is a faulty machine in the original population based on the operation and maintenance data;

[0042] If it exists, the fault operation and maintenance data corresponding to the faulty machine is obtained, and based on the fault operation and maintenance data, the fault location of the faulty machine is determined, and the machine production parts are matched with the production process of the product to be produced to obtain a production operation pairing group;

[0043] Based on the fault location and combined with the production operation pairing group, the abnormal production process is determined;

[0044] And obtain the production correlation relationship between the faulty machine and the remaining machines, and based on the production relationship, determine whether the faulty machine can perform abnormal production process jump;

[0045] If possible, obtain the assisting machine corresponding to the abnormal production process jump, generate a process skipping label based on the abnormal production process and the assisting machine, and add it to the corresponding machine in the optimal population arrangement sequence;

[0046] Otherwise, a machine fault elimination instruction is generated to eliminate the faulty machine from the optimal population arrangement sequence, and based on the total number of eliminated faulty machines, the population size is adaptively adjusted and the execution continues;

[0047] If it does not exist, go directly to step 1.

[0048] Preferably, in a method for solving the flexible workshop scheduling problem based on the improved golden jackal optimization algorithm, iterative processing is performed according to the position update result, including:

[0049] When a process skip tag is detected in the machine selection part, the abnormal production process corresponding to the current iteration selected machine is automatically replaced by the assisting process corresponding to the assisting machine.

[0050] Compared with the prior art, the present invention has at least the following beneficial effects:

[0051] The present invention generates an initialized population by combining GLS initialization with RI initialization, calculates the fitness values ​​of individuals in the initialized population, and initializes the original population by combining random inversion with combined heuristic rules, which is beneficial to improving the generation diversity and high quality of the initialized population and providing a basis for the golden jackal to improve the diversity; then the fitness values ​​of the individuals are compared, and according to the comparison results, the male golden jackal (i.e., the optimal individual) and the female golden jackal (i.e., the suboptimal individual) are determined to provide an initial basis for improving the operation of the golden jackal optimization algorithm; then the current maximum number of iterations is set to 0, the positions of the machine selection part and the process sorting part are updated respectively, and iterative processing is performed according to the position update results, and based on the iterative processing results, the current optimal solution is updated; and the judgment Whether the current iteration meets the iteration termination condition, if it meets the iteration termination condition, then output the optimal solution, otherwise execute step 3 for the next iteration; achieve the goal of using multiple strategies to improve the golden jackal optimization algorithm to completely solve the flexible workshop scheduling problem, and at the same time, based on the new nonlinear convergence factor, improve the calculation of the prey loss energy of the golden jackal in the prey search process in the golden jackal optimization algorithm, which is beneficial to improve the global search ability of the golden jackal optimization algorithm; based on the golden sine algorithm, improve the position update formula of the golden jackal in the prey hunting process, which is beneficial to improve the local search ability of the golden jackal optimization algorithm, and search for the optimal solution more efficiently. While improving the local search ability of the golden jackal optimization algorithm, the global search ability is improved to ensure that the best solution that meets the global deployment of the workshop is provided for the flexible workshop scheduling problem.

[0052] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.

[0053] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0055] Figure 1It is a flow chart of a method for solving the flexible workshop scheduling problem based on the improved golden jackal optimization algorithm of the present invention;

[0056] Figure 2 This is a flow chart of step 3 of a method for solving a flexible shop scheduling problem based on an improved golden jackal optimization algorithm of the present invention;

[0057] Figure 3 Flowchart for the conversion from continuous space to discrete space;

[0058] Figure 4 Flowchart for the conversion from discrete space to continuous space;

[0059] Figure 5 A schematic diagram of the variable neighborhood search strategy of the present invention;

[0060] Figure 6 The present invention is a flowchart of step 1 of a method for solving a flexible shop scheduling problem based on an improved golden jackal optimization algorithm. DETAILED DESCRIPTION

[0061] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0062] Embodiment 1:

[0063] The present invention provides a method for solving the flexible workshop scheduling problem based on an improved golden jackal optimization algorithm, such as Figure 1 As shown, including:

[0064] Step 1: Generate an initialization population based on the combination of GLS initialization and RI initialization, and calculate the fitness values ​​of individuals in the initialization population;

[0065] Step 2: Compare the fitness values ​​of the individuals, and determine the male golden jackal and the female golden jackal based on the comparison results;

[0066] Step 3: Set the current maximum number of iterations to 0, update the positions of the machine selection part and the process sorting part respectively, perform iterative processing according to the position update results, and update the current optimal solution based on the iterative processing results;

[0067] Step 4: Determine whether the current iteration meets the iteration termination condition. If so, output the optimal solution. Otherwise, execute step 3 for the next iteration.

[0068] Before executing step 1, the following steps are included:

[0069] Get relevant iteration parameter setting information;

[0070] The relevant iteration parameters include population size, maximum number of iterations, crossover probability factor, mutation probability factor and initialization rule fusion ratio;

[0071] Before executing step 1, it includes: improving the golden jackal optimization algorithm based on a new nonlinear convergence factor and a golden sine algorithm.

[0072] In this embodiment, GLS initialization refers to the initialization rule obtained by fusing the global selection (GS) rule, the local selection (LS) rule, and the random selection (RS) rule according to a certain fusion ratio, wherein the sum of the fused person ratios corresponding to the GS rule, the LS rule, and the RS rule is 1.

[0073] GS starts from a global perspective and ensures that the machine with the shortest processing time is given priority to balance the load of the entire processing machine; LS starts from a local perspective and ensures that the process with the shortest processing time in a workpiece is given priority to balance the load on each independent machine; RS uses random initialization to disperse the initialization population throughout the solution space.

[0074] In this embodiment, RI initialization refers to the initialization rule obtained by fusing the two rules of Random Selection (RS) and Invert Selection (IS) according to a certain fusion ratio, wherein the sum of the fusion ratios corresponding to the IS rule and the RS rule is 1.

[0075] RS randomly shuffles the process codes generated in the encoding method, which helps to improve the diversity of the initial population; IS reverses the order of the process codes generated in the encoding method, generates new process plans, and enhances the breadth of the initial population.

[0076] The present invention adopts two-segment integer coding, corresponding to the machine selection part and the process sorting part respectively, and the length of each segment of coding is the total number of workpieces. The first part is the machine selection part code (MS), which is used to determine the processing machine of each process on the workpiece, and the second part is the process sorting code (OS), which is used to determine the order of processing the workpieces:

[0077] The MS part is updated by using a random crossover strategy on the initial population, that is, traversing the entire population, selecting several genes on two individuals in the population for exchange operations under the crossover probability, otherwise the individuals of the initial population are used;

[0078] The OS part adopts a mechanism of mutual conversion between continuous space and discrete space coding, including two steps: conversion from continuous space to discrete space and conversion from discrete space to continuous space, as follows:

[0079] Step A: Conversion from continuous space to discrete space. The specific steps are as follows: Figure 3 As shown:

[0080] First, randomly select values ​​within the specified range [-m, m] and correspond them to each process one by one. Then, use the ROV rule (Ranked Order Value, ascending sorting rule) to assign values ​​to each process, that is, sort the generated random numbers in ascending order from small to large and assign element numbers (i.e., ROV values) and element sequence numbers. Finally, rearrange the ROV values ​​and element sequence codes according to the original randomly generated random numbers to obtain a new process sorting scheme, where m represents the total number of machines available for selection in the machine selection part;

[0081] Step B: Conversion from discrete space to continuous space. The specific steps are as follows: Figure 4 As shown:

[0082] First, randomly select values ​​within the specified range [-m, m] and correspond to each process. Then, sort the random numbers in ascending order and assign ROV values ​​and element sequence numbers. Finally, reorder the ROV values ​​according to the process sorting scheme to obtain a continuous space vector.

[0083] Correspondingly, the present invention adopts a two-stage decoding method, and its decoding strategy includes:

[0084] For the decoding of the machine part, the machine part code is read from left to right and converted into a machine sequence matrix and a time sequence matrix. For the decoding of the process part, the process insertion method is adopted. On the basis of not changing the scheduled process, the subsequent process is inserted into the idle time period of the corresponding machine as much as possible, thus shortening the maximum completion time of the corresponding machine.

[0085] The present invention adopts a discrete integer code consisting of two parts to realize the combination of the improved golden jackal optimization algorithm and the flexible workshop scheduling problem.

[0086] In this embodiment, the improvement of the golden jackal optimization algorithm based on the new nonlinear convergence factor refers to improving the calculation of the prey loss energy of the golden jackal in the prey search process in the golden jackal optimization algorithm based on the new nonlinear convergence factor.

[0087] Among them, the new nonlinear convergence factor is calculated as follows:

[0088]

[0089] Among them, E 1represents the new nonlinear convergence factor (i.e., prey loss energy); t represents the current maximum number of iterations of the population; T represents the maximum number of iterations of the population.

[0090] The improvement of the golden jackal optimization algorithm based on the golden sine algorithm refers to improving the position update formula of the golden jackal in the process of prey hunting based on the golden sine algorithm, and improving the position update formula as follows:

[0091] Y 1 (t)=|sin(r 1 )*Prey(t)+sin(r 1 )*r 2 *|x 1 *Y M (t)-x 2 *Prey(t)|

[0092] Y 2 (t)=|sin(r 1 )*Prey(t)+sin(r 1 )*r 2 *|x 1 *Y FM (t)-x 2 *Prey(t)|

[0093] Among them, Y 1 (t) represents the updated position Y of the male golden jackal 2 (t) indicates the update position of the female golden jackal; r 1 is a random number, ranging from [0, 2π]; r 2 is a random number in [0,π]; x 1 and x 2 The coefficient x obtained by introducing the golden section 1 =π*(1-μ)+(-π*μ), x 2 =-π*(1-μ)+(π*μ); μ represents the golden section number, Prey(t) represents the prey position vector; Y M (t) indicates the current position of the male golden jackal; Y FM (t) indicates the current position of the female golden jackal.

[0094] The beneficial effects of the above technical solution are as follows: the present invention generates an initialized population by combining GLS initialization with RI initialization, calculates the fitness values ​​of individuals in the initialized population, and initializes the original population by combining random inversion with combined heuristic rules, which is beneficial to improving the generation diversity and high quality of the initialized population, and provides a basis for the golden jackal to improve the diversity; then the fitness values ​​of the individuals are compared, and based on the comparison results, the male golden jackal (i.e., the optimal individual) and the female golden jackal (i.e., the suboptimal individual) are determined to provide an initial basis for improving the operation of the golden jackal optimization algorithm; then the current maximum number of iterations is set to 0, the positions of the machine selection part and the process sorting part are updated respectively, and iterative processing is performed according to the position update results, and based on the iterative processing results, the current optimal solution is updated; the current iteration is judged Whether the iteration termination condition is met, if the iteration termination condition is met, the optimal solution is output, otherwise step 3 is executed for the next iteration; the goal of using multiple strategies to improve the golden jackal optimization algorithm to completely solve the flexible workshop scheduling problem is achieved, and at the same time, based on the new nonlinear convergence factor, the calculation of the prey loss energy of the golden jackal in the prey search process in the golden jackal optimization algorithm is improved, which is beneficial to improve the global search ability of the golden jackal optimization algorithm; improving the golden jackal optimization algorithm based on the golden sine algorithm means improving the position update formula of the golden jackal in the prey hunting process based on the golden sine algorithm, which is beneficial to improve the local search ability of the golden jackal optimization algorithm, and more efficiently search for the optimal solution. While improving the local search ability of the golden jackal optimization algorithm, the global search ability is improved to ensure that the best solution that meets the global deployment of the workshop is provided for the flexible workshop scheduling problem.

[0095] Embodiment 2:

[0096] Based on Example 1, step 3, as Figure 2 As shown, specifically including:

[0097] Step 301: Calculate the fitness value of the individual after the position is updated according to the decoding strategy. If the fitness value is greater than the current individual, replace the current individual based on the individual after the position is updated.

[0098] Step 302: Perform a crossover mutation operation on the replaced individual, and calculate a new solution of the crossover mutation operation according to a decoding strategy. If the new solution is better than the solution corresponding to the current individual, the new solution is accepted as the optimal solution.

[0099] Step 303: Execute the variable neighborhood search strategy on the optimal solution, generate a new neighborhood solution and evaluate it. If the new neighborhood solution is better than the current optimal solution, the current optimal solution is updated.

[0100] In this embodiment, for the crossover operation, the machine selection part selects random reserved crossover under the crossover probability, and the process sorting part selects the improved reserved process sequence crossover operation under the crossover probability. For the mutation operation, the machine selection part adopts the single point mutation method under the mutation probability. And the process sorting part adopts the nearest neighbor mutation method under the mutation probability.

[0101] In this embodiment, the variable neighborhood search strategy refers to changing the order of different processes on the critical path without changing the selection of the corresponding critical process processing machine. The present invention designs different neighborhood structures according to the position and number of key blocks (multiple non-interval key processes on the same machine constitute a key block), specifically Figure 5 As shown:

[0102] 1) If the starting key block includes more than two processes, only the two processes connected at the end of the block are exchanged. If the two connected processes belong to the same workpiece, they are not exchanged;

[0103] 2) If the internal key block contains three or more processes, swap the block head with the second to last process, and swap the block tail with the second process. ;

[0104] 3) If the last key block contains more than two processes, only the two processes connected at the beginning of the block are exchanged. If the two connected processes belong to the same workpiece, they are not exchanged.

[0105] Among them, the critical path refers to the longest path from the first process to the last process of the machine in a FJSP feasible scheduling plan without any idle time in between;

[0106] Each process on the critical path is called a critical process.

[0107] The beneficial effects of the above technical solution are as follows: first, the fitness value of the individual after the position is updated is calculated according to the decoding strategy. If the fitness value is greater than the current individual, the current individual is replaced based on the individual after the position is updated; then a crossover mutation operation is performed on the replaced individual, and a new solution of the crossover mutation operation is calculated according to the decoding strategy. If the new solution is better than the solution corresponding to the current individual, the new solution is accepted as the optimal solution; finally, a variable neighborhood search strategy is performed on the optimal solution to generate a new neighborhood solution and evaluate it. If the new neighborhood solution is better than the current optimal solution, the current optimal solution is updated, and an iterative process is performed to realize automatic selection of the optimal solution. Genetic operations such as random retention crossover and single-point mutation are introduced during the iterative execution process, which is beneficial to improving the local search capability of the golden jackal optimization algorithm; a variable neighborhood operation strategy based on a critical path is used to deeply mine the optimal individual after each iteration, which is beneficial to expanding the golden jackal search space, avoiding excessive local optimization, and providing a solution strategy that is more in line with the global deployment of the workshop for the scheduling problem of the flexible workshop.

[0108] Embodiment 3:

[0109] On the basis of Example 2, the fitness value of the individual after the position is updated is calculated according to the decoding strategy. If the fitness value is greater than the current individual, the current individual is replaced based on the individual after the position is updated. In the process, the Cauchy-Gaussian mutation strategy is used to randomly perturb the optimal individual generated in each iteration.

[0110] In this embodiment, the Cauchy-Gaussian variation formula is as follows:

[0111] Y new (t) = Y(t)[1+λ 1 cauchy(0,1)+(1-λ 1 )Gauss(0,1)]

[0112]

[0113] Where Y(t) represents the optimal individual produced after each iteration, λ 1 represents the adaptive adjustment parameter, cauchy(0,1) and Gauss(0,1) represent random numbers that obey the Cauchy distribution and Gaussian distribution respectively.

[0114] The beneficial effects of the above technical solution are as follows: when calculating the fitness value of the individual after the position update according to the decoding strategy, if the fitness value is greater than the current individual, the current individual is replaced based on the individual after the position update. In the process, the Cauchy-Gaussian mutation strategy is used to randomly perturb the optimal individual generated in each iteration. The optimal individual is perturbed by the new Cauchy-Gaussian mutation strategy, which effectively avoids local stagnation of the golden jackal optimization algorithm.

[0115] Embodiment 4:

[0116] On the basis of Example 1, step 1, such as Figure 6 As shown, specifically including:

[0117] Step 101: respectively obtaining machine selection fusion ratios of multiple first target selection rules included in GLS initialization;

[0118] Respectively obtain the process ranking fusion ratios of multiple second target selection rules included in RI initialization;

[0119] Step 102: According to the machine selection fusion ratio, the first target selection rule is fused to generate a GLS initialization rule, and the machine selection part of the original population is initialized based on the GLS initialization rule to obtain a machine selection part initialization result;

[0120] Synchronously, according to the process sorting fusion ratio, the second target selection rule is fused to generate the RI initialization rule, and the process sorting part of the original population is initialized based on the RI initialization rule to obtain the initialization result of the process sorting part;

[0121] Step 103: Based on the initialization results of the machine selection part and the initialization results of the process sorting part, an initialization population is obtained;

[0122] Step 104: Based on the genetic algorithm, calculate the fitness values ​​of individuals in the initialized population.

[0123] The beneficial effects of the above technical solution: the present invention adopts a fusion of multiple selection rules to generate an initialization strategy that combines random inversion and combined heuristic rules, and uses this initialization strategy to initialize the machine selection part and the process sorting part with different rules respectively, which is conducive to generating a high-quality and diverse initial population and improving the solution efficiency of the subsequent golden jackal optimization algorithm.

[0124] Embodiment 5:

[0125] On the basis of Example 1, after executing step 1, the method includes:

[0126] Based on the initial population data, the iteration setting parameters are intelligently modified, including:

[0127] Based on the population size setting information, determine the number of production-adaptive machines for the products to be produced and the types of products to be produced;

[0128] Based on the type of product to be produced, obtain the production process of the product to be produced, determine the production process of the product to be produced, and the sequential relationship between the processes;

[0129] According to the number of production adaptation machines and the sequential association relationship, obtain the historical scheduling processing data corresponding to the machines of similar historical production products;

[0130] Based on historical scheduling processing data, predict the best scheduling strategy for products to be produced;

[0131] According to the optimal scheduling strategy, the optimal solution position is predicted for the current initialized population, and the number of iteration terminations is determined based on the prediction results;

[0132] and comparing the number of iteration terminations with the maximum number of iterations in the relevant iteration parameter setting information to obtain a setting error degree, and when the setting error degree is greater than a preset value, modifying the maximum number of iterations based on the number of iteration terminations;

[0133] Otherwise, the number of iteration terminations remains unchanged.

[0134] In this embodiment, the products to be produced refer to the products to be produced corresponding to the flexible workshop scheduling problem.

[0135] In this embodiment, the production-compatible machine refers to a machine in a workshop that can complete the production of the product to be produced.

[0136] In this embodiment, the number of iteration terminations refers to the predicted optimal number of iterations of the current initialized population.

[0137] In this embodiment, the setting error degree refers to the ratio of the difference between the number of iteration termination times and the maximum number of iterations to the number of iteration termination times.

[0138] The beneficial effects of the above technical solution are as follows: after obtaining the initialization population, the present invention determines the number of production adaptation machines for the products to be produced and the types of products to be produced based on the population size setting information;

[0139] Based on the type of product to be produced, obtain the production process of the product to be produced, determine the production process of the product to be produced, and the sequential relationship between the processes;

[0140] According to the number of production adaptation machines and the sequential association relationship, obtain the historical scheduling processing data corresponding to the machines of similar historical production products;

[0141] Based on historical scheduling processing data, predict the best scheduling strategy for products to be produced;

[0142] According to the optimal scheduling strategy, the optimal solution position is predicted for the current initialized population, and the number of iteration terminations is determined based on the prediction results; and the number of iteration terminations is compared with the maximum number of iterations in the relevant iteration parameter setting information to obtain the setting error degree. When the setting error degree is greater than the preset value, the maximum number of iterations is modified based on the number of iteration terminations; otherwise, the number of iteration terminations is kept unchanged. This realizes the intelligent modification of the number of iterations according to the population data of the golden jackal optimization algorithm, which ensures the optimal solution to the flexible workshop scheduling problem while reducing the number of invalid iterations, effectively improving the efficiency of solving the scheduling problem.

[0143] Embodiment 6:

[0144] On the basis of Example 5, based on the historical scheduling processing data, the optimal scheduling strategy for the product to be produced is predicted, including:

[0145] Based on historical scheduling processing data, multiple optimal selection rankings are obtained, and multiple different production scheduling sequences are determined based on the optimal selection rankings;

[0146] Compare multiple different production scheduling sequences to obtain the same scheduling subsequences and different scheduling subsequences;

[0147] Extract scheduling features for different scheduling subsequences respectively, obtain scheduling similar features and compare the features, and schedule similar features;

[0148] Based on the similarity characteristics of the scheduling and the local process characteristics corresponding to the same scheduling subsequence, the sorting order of the initialized population is adjusted to obtain the optimal scheduling strategy.

[0149] The beneficial effects of the above technical solution are as follows: the present invention obtains multiple optimal selection rankings based on historical scheduling processing data, and determines multiple different production scheduling sequences based on the optimal selection rankings; compares multiple different production scheduling sequences to obtain the same scheduling subsequences and different scheduling subsequences; extracts scheduling features from different scheduling subsequences respectively to obtain scheduling similar features and compares the features to obtain scheduling similar features; based on the scheduling similar features and the local process features corresponding to the same scheduling subsequences, adjusts the sorting order of the initialized population to obtain the optimal scheduling strategy, realizes the prediction of the optimal scheduling strategy for the products to be produced through historical scheduling processing data, and provides a basis for intelligent modification of the number of iterations.

[0150] Embodiment 7:

[0151] On the basis of Example 1, before executing step 1, the method includes:

[0152] After the user inputs the original population information, the operation and maintenance data of each machine in the original population is obtained, and based on the operation and maintenance data, it is determined whether there is a faulty machine in the original population;

[0153] If it exists, the fault operation and maintenance data corresponding to the faulty machine is obtained, and based on the fault operation and maintenance data, the fault location of the faulty machine is determined, and the machine production parts are matched with the production process of the product to be produced to obtain a production operation pairing group;

[0154] Based on the fault location and combined with the production operation pairing group, the abnormal production process is determined;

[0155] And obtain the production correlation relationship between the faulty machine and the remaining machines, and based on the production relationship, determine whether the faulty machine can perform abnormal production process jump;

[0156] If possible, obtain the assisting machine corresponding to the abnormal production process jump, generate a process skipping label based on the abnormal production process and the assisting machine, and add it to the corresponding machine in the optimal population arrangement sequence;

[0157] Otherwise, a machine fault elimination instruction is generated to eliminate the faulty machine from the optimal population arrangement sequence, and based on the total number of eliminated faulty machines, the population size is adaptively adjusted and the execution continues;

[0158] If it does not exist, go directly to step 1.

[0159] In this embodiment, the operation pairing group refers to a pairing group formed by one-to-one correspondence between the components processed and produced by each part of the machine and the process of the product to be produced.

[0160] In this embodiment, an abnormal production process refers to a process that cannot be completed by a faulty machine.

[0161] In this embodiment, the remaining machines refer to other machines in the workshop except the currently faulty machine.

[0162] In this embodiment, the generation association relationship refers to whether the products generated by different machines can be exchanged with each other at a certain process, or whether one machine can take over the current production results of another machine to continue the remaining processes, or can produce part of the remaining processes and then transfer them back to the original machine production line.

[0163] In this embodiment, the assisting machine refers to other machines that have a generated association relationship with the faulty machine.

[0164] In this embodiment, the process skipping label refers to a label carried by a faulty machine that needs to skip an abnormal production process and assist the machine to complete the abnormal production process.

[0165] The beneficial effects of the above technical solution are as follows: the present invention screens all machines in the original population before executing the steps, proposes faulty machines that cannot be used, avoids the situation where the optimal solution to the lucky shop scheduling problem obtained by the golden jackal optimization algorithm is an invalid solution, and retains machines that can make some processes fail, which is conducive to improving population diversity.

[0166] Embodiment 8:

[0167] Based on Embodiment 7, iterative processing is performed according to the location update result, including:

[0168] When a process skip tag is detected in the machine selection part, the abnormal production process corresponding to the current iteration selected machine is automatically replaced by the assisting process corresponding to the assisting machine.

[0169] The beneficial effects of the above technical solution are as follows: when the machine selection part detects a process skip label, the present invention automatically replaces the abnormal production process corresponding to the currently iteratively selected machine with the assisting process corresponding to the assisting machine. While improving the diversity of the golden jackal population, it is beneficial to improve the selectable range of solutions to the flexible workshop scheduling problem and the utilization rate of machines in the workshop, thereby improving the workshop production capacity.

[0170] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for solving the flexible shop scheduling problem based on an improved golden jackal optimization algorithm, characterized in that: include: Step 1: Generate an initialization population based on the combination of GLS initialization and RI initialization, and calculate the fitness values ​​of individuals in the initialization population; Among them, RI initialization refers to the initialization rule obtained by fusing the two rules of random process selection and reverse inversion selection according to the fusion ratio; Step 2: Compare the fitness values ​​of the individuals, and determine the male golden jackal and the female golden jackal based on the comparison results; Step 3: Set the current maximum number of iterations to 0, update the positions of the machine selection part and the process sorting part respectively, perform iterative processing according to the position update results, and update the current optimal solution based on the iterative processing results; Step 4: Determine whether the current iteration meets the iteration termination condition. If so, output the optimal solution. Otherwise, execute step 3 for the next iteration. Before executing step 1, include: Get relevant iteration parameter setting information; The relevant iteration parameters include population size, maximum number of iterations, crossover probability factor, mutation probability factor and initialization rule fusion ratio; The golden jackal optimization algorithm is improved based on the new nonlinear convergence factor and golden sine algorithm; Among them, the new nonlinear convergence factor is calculated as follows: in, represents the new nonlinear convergence factor, i.e., prey loss energy; t represents the maximum number of iterations of the current population; T represents the maximum number of iterations of the population; After executing step 1, including: Based on the initial population data, the iteration setting parameters are intelligently modified, including: Based on the population size setting information, determine the number of production-adaptive machines for the products to be produced and the types of products to be produced; Based on the type of product to be produced, obtain the production process of the product to be produced, determine the production process of the product to be produced, and the sequential association relationship between the processes; According to the number of production adaptation machines and the sequential association relationship, obtain the historical scheduling processing data corresponding to the machines of similar historical production products; Based on historical scheduling processing data, predict the best scheduling strategy for products to be produced; According to the optimal scheduling strategy, the optimal solution position is predicted for the current initialized population, and the number of iteration terminations is determined based on the prediction results; and comparing the number of iteration terminations with the maximum number of iterations in the relevant iteration parameter setting information to obtain a setting error degree, and when the setting error degree is greater than a preset value, modifying the maximum number of iterations based on the number of iteration terminations; Otherwise, keep the number of iteration terminations unchanged; Wherein, step 3 specifically includes: Calculate the fitness value of the individual after the position is updated according to the decoding strategy, and if the fitness value is greater than the current individual, replace the current individual based on the individual after the position is updated; Perform a crossover mutation operation on the replaced individual, and calculate a new solution of the crossover mutation operation according to the decoding strategy. If the new solution is better than the solution corresponding to the current individual, the new solution is accepted as the optimal solution; Execute a variable neighborhood search strategy on the optimal solution, generate a new neighborhood solution and evaluate it. If the new neighborhood solution is better than the current optimal solution, update the current optimal solution. Among them, the neighborhood structure of the variable neighborhood search strategy specifically includes: 1) If the starting key block includes more than two processes, only the two processes connected at the end of the block are exchanged. If the two connected processes belong to the same workpiece, they are not exchanged; 2) If the internal key block contains three or more processes, swap the block head with the second-to-last process, and swap the block tail with the second process; 3) If the last key block contains more than two processes, only the two processes connected at the beginning of the block are exchanged. If the two connected processes belong to the same workpiece, they are not exchanged; When calculating the fitness value of the individual after the position update according to the decoding strategy, if the fitness value is greater than the current individual, the Cauchy-Gauss mutation strategy is used to randomly perturb the optimal individual generated in each iteration in the process of replacing the current individual based on the individual after the position update; Among them, the Cauchy-Gaussian variation formula is as follows: in represents the optimal individual produced after each iteration, represents the adaptive adjustment parameter, cauchy(0,1) and Gauss(0,1) represent random numbers that obey the Cauchy distribution and Gaussian distribution respectively.

2. The method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm according to claim 1 is characterized in that: Step 1 specifically includes: Respectively obtain the machine selection fusion ratios of multiple first target selection rules included in the GLS initialization; Respectively obtain the process ranking fusion ratios of multiple second target selection rules included in RI initialization; According to the machine selection fusion ratio, the first target selection rule is fused to generate a GLS initialization rule, and the machine selection part of the original population is initialized based on the GLS initialization rule to obtain the machine selection part initialization result; According to the process sorting fusion ratio, the second target selection rules are fused to generate RI initialization rules, and the process sorting part of the original population is initialized based on the RI initialization rules to obtain the initialization result of the process sorting part; Based on the initialization results of the machine selection part and the initialization results of the process sorting part, an initialization population is obtained; Based on the genetic algorithm, the fitness values ​​of individuals in the initialized population are calculated.

3. The method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm according to claim 1 is characterized in that: Based on historical scheduling processing data, the optimal scheduling strategy for the products to be produced is predicted, including: Based on historical scheduling processing data, multiple optimal selection rankings are obtained, and multiple different production scheduling sequences are determined based on the optimal selection rankings; Compare multiple different production scheduling sequences to obtain the same scheduling subsequences and different scheduling subsequences; Extract scheduling features for different scheduling subsequences respectively, obtain scheduling similar features and compare the features, and schedule similar features; Based on the similarity characteristics of the scheduling and the local process characteristics corresponding to the same scheduling subsequence, the sorting order of the initialized population is adjusted to obtain the optimal scheduling strategy.

4. The method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm according to claim 1 is characterized in that: Before executing step 1, it includes: after the user inputs the original population information, obtaining the operation and maintenance data of each machine in the original population, and judging whether there is a faulty machine in the original population based on the operation and maintenance data; If it exists, the fault operation and maintenance data corresponding to the faulty machine is obtained, and based on the fault operation and maintenance data, the fault location of the faulty machine is determined, and the machine production parts are matched with the production process of the product to be produced to obtain a production operation pairing group; Based on the fault location and combined with the production operation pairing group, the abnormal production process is determined; And obtain the production correlation relationship between the faulty machine and the remaining machines, and based on the production correlation relationship, determine whether the faulty machine can perform abnormal production process jump; If possible, obtain the assisting machine corresponding to the abnormal production process jump, generate a process skipping label based on the abnormal production process and the assisting machine, and add it to the corresponding machine in the optimal population arrangement sequence; Otherwise, a machine fault elimination instruction is generated to eliminate the faulty machine from the optimal population arrangement sequence, and based on the total number of eliminated faulty machines, the population size is adaptively adjusted and the execution continues; If it does not exist, go directly to step 1.

5. The method for solving the flexible shop scheduling problem based on the improved golden jackal optimization algorithm according to claim 4 is characterized in that: Iterate based on the location update result, including: When a process skip tag is detected in the machine selection part, the abnormal production process corresponding to the current iteration selected machine is automatically replaced by the assisting process corresponding to the assisting machine.

Images