Design And Optimization Of Improved Adaptive Genetic Algorithm For Flexible Job Shop Scheduling Problem In G Company

In recent years,with the improvement of productivity,the demand of consumers has become more and more individualized,which has driven the manufacturing orders to appear more and more “small amount and multiple batches”.This has raised the flexibility of the production workshop.Requirements,production scheduling needs to face a larger problem scale.The research and development of flexible job-shop scheduling problem(FJSP)meets this requirement to a large extent,and many solutions can quickly give a satisfactory satisfaction for large-scale scheduling problems' Scheduling plan.The flexible job shop production scheduling problem is a basic problem model with wide applicability and good scalability.However,with the continuous improvement of the mechanization and automation of the manufacturing industry,many new production systems have been added to the workshops,resulting in a new variant model for the production scheduling problem in the job shop.The corresponding scheduling solutions need to be more suitable according to the specific conditions of the workshop.Therefore,this paper takes G company's flexible operation workshop production and processing as the research goal,and analyzes two complexities compared with the typical flexible job shop production scheduling problem-Automated-Guided Vehicle(AGV)The transportation time and the number of dynamic machines,and the point-to-face design is a suitable scheduling method for this type of job shop.The genetic algorithm has the advantages of fast solution speed,strong global search ability and good robustness.It is suitable for solving complex problems with large background and large scale.Therefore,the genetic algorithm is the basic algorithm of the solution and further improved and optimized.Because the production scheduling problem of G company's flexible operation shop is naturally divided into “process segment” and “machine segment”,the coding method is designed as two-stage coding,which is designed to match the POX(precedence operation crossover)crossover operator and multi-point randomization.The mutation operator,the reciprocal of the maximum process time is the fitness function and the roulette selection operator.In the analysis of Kacem example test and G company scheduling problem solving scheme,the problems of unbalanced machine load and slow convergence of the algorithm are found,which makes the optimal solution of the problem unsatisfactory.For this purpose,a greedy insertion decoding strategy,a prior machine selection mechanism based on normal distribution,a machine load balancing auxiliary objective function design,a machine load compensation selection mechanism and an evolution operator are designed.And a satisfactory solution is obtained in the example test and the solution of the G company scheduling problem,but the example still cannot find the optimal solution stably.In order to obtain a further optimization solution,improve the convergence speed and convergence stability of the algorithm,and simplify the parameter setting of the algorithm,this paper adds the adaptive genetic algorithm operator design on the improved genetic algorithm.The construction of adaptive function combines logistic function and Gaussian distribution function.In addition to the commonly used iteration number,the similarity parameters and population stagnation algebra parameters based on elite individual coding are designed.Not only adaptive functions are designed for the probability and length of crossover,mutation and evolution.In order to adjust the population search ability,the adaptive operator adaptive function is also designed innovatively.