Research On Flexible Job Shop Scheduling Problem Based On Swarm Intelligence Algorithm

In modern manufacturing,indicators such as efficiency,energy conservation,reliability,stability,and green low-carbon are key indicators pursued by enterprises in the production process.Developing excellent scheduling plans is one of the important methods to achieve these indicators.Optimizing scheduling plans can maximize the utilization of production resources,improve production efficiency,and enhance enterprise competitiveness.In the field of workshop scheduling,the classic job scheduling problem is transitioning to a workshop scheduling problem with characteristics such as multi-objective,flexibility,and distribution.This kind of problem is a complex combinatorial optimization problem,and the difficulty of the problem will increase with the increase of the scale,which is consistent with the application scenario of intelligent manufacturing.Swarm intelligence algorithm shows good performance in solving combinatorial optimization problems.It has high search efficiency,good stability and strong adaptability to similar models when optimizing objectives.This article mainly studies three different types of flexible job shop scheduling problems(FJSP),solves them by improving the basic algorithm,and verifies the effectiveness of the algorithm.The main content is as follows:(1)Adopting an improved artificial bee colony algorithm to solve the flexible job shop scheduling problem with the goal of minimizing the maximum completion time,a case cleaning strategy is proposed to reduce the solution space by combining the discrete characteristics of job shop scheduling with the crossover operator in genetic algorithm to assist bee colony individuals in optimization.This algorithm first uses double-layer encoding to establish the connection between processing information and individuals,and creates a population by mixing multiple initialization strategies to improve individual quality;Then,design multiple processes and neighborhood structures of machines to enhance the local search ability of bee colony individuals;Next,considering the impact of collaborative goals,design new individual selection methods for crossover,and design a variable neighborhood search strategy to improve local optimization ability;Finally,multiple sets of experiments were conducted to verify the algorithm’s good convergence performance.(2)Using improved artificial bee colony algorithm to solve multi-objective flexible job shop scheduling problem.This algorithm first uses double-layer encoding to represent workshop information,improves the judgment conditions of non dominated solutions,and proposes an update strategy to accelerate the exploration of honey source potential;Then,different crossover operators and mutation strategies are selected to update individuals based on their reproductive ability;Next,two selection strategies are redefined to improve the competitive level of individuals,and two learning mechanism based strategies are used to replace weaker individuals and improve population vitality;Finally,two sets of numerical examples and multiple sets of workshop examples were used to validate the effectiveness of the improved strategy,which demonstrated good solving performance in the comparison algorithm.(3)Using an improved grey wolf optimization algorithm to solve the fuzzy flexible job shop scheduling problem.Firstly,the algorithm uses double coding to discretization the gray wolf individuals,and designs an initialization strategy based on the contribution degree of the super volume index to improve the population diversity;Then,reinforcement learning methods are used to determine global and local search parameters,and two crossover operators are improved to assist individuals in evolving under different update modes;Next,use two-level variable neighborhood and four replacement strategies to improve local search ability;Finally,separation experiments were conducted on multiple examples to verify the effectiveness of the improved strategy.Comparative experiments were conducted on most examples,proving that the algorithm performs better than the comparative algorithm and has good convergence and distribution.