Optimization And Scheduling Mode Of The Flexible Flow-shop With Multiprocessor Tasks

With the concept of "Industrial Manufacturing 4.0",manufacturing industry is facing both opportunities and increasingly fierce competition.How to improve the production efficiency and service quality of manufacturing industry has become the key point.As the core issue of manufacturing enterprise operation,production scheduling has become a hot research topic in academia and engineering circles.The Flexible Flow-shop Scheduling Problem(FFSP)was originally refined from petroleum and chemical industries,so it has a strong industrial application background.FFSP structure exists in most manufacturing enterprises(steel,pharmaceutical,chemical),container handling,assembly,transportation and other systems.The FFSP with multi-processor tasks studied in this paper is a combination of multi-processor tasks and FFSP environment.Simplification of one or more feature constraints in the problem may constitute other typical scheduling models.Therefore,the research is generalized and challenging.In this paper,FFSP with multi-processor tasks is studied,with considered the actual production constraints such as time window,dynamic arrival and transportation time of jobs.In the past literature,the objective value of FFSP with multi-processor tasks is to minimize the maximum completion time,which measures the total completion time of job processing.The smaller the value,the higher the production efficiency.Literatures that aim at minimizing the total weighted completion time have also attracted much attention from scholars.Therefore,in order to minimize the maximum completion time for FFSP with multiprocessor tasks with transport time,GA& IGP(Genetic Algorithm &Iterative Greedy Procedure)is used to solve the model.Then,the target value of FFSP with time window and multi-processor task is taken to minimize the total weighted completion time,which is solved by HGA(Hybrid Genetic Algorithm).The algorithm adopted in this paper is mainly based on the greedy algorithm and the idea of adaptive improvement to improve the genetic algorithm.Firstly,based on the complexity of the research problem,the job not only needs to determine the order of the job,but also needs to arrange a sufficient number of machines for processing.Therefore,a two-dimensional coding scheme based on job-machine is designed,which uses JAP-RSIM job scheduling mechanism to generate the initial feasible solution.Then,a new solution updating process based on job-machine synchronous crossover and variation is carried out.After calculating and judging the fitness value of the new solution,a new demodulation,adjustment and reconstruction operation of greedy algorithm is carried out,and a GA-IGP fusion optimization strategy is formed.On the basis of GA& IGP program,the idea of self-adaptive improvement is introduced to improve the mutation probability and cross-improvement adaptively,and a genetic algorithm-based HGA algorithm is formed.Finally,the simulation experiments are used to test several groups of different scale problems.The GA& IGP is compared with GA,IGP and GA& IGP in solving FFSP with multi-processor task with the same constraints.The GA&IGP has the best solution quality and is suitable for solving large-scale problems.Furthermore,three improved genetic algorithms,GA& IGP and HGA,are compared in solving FFSP with multi-processor tasks with the same constraints.The quality of HGA is better than GA&IGP and GA in solving FFSP with the same constraints.It is proved that the introduction of greedy algorithm to improve genetic algorithm can avoid the shortcoming of premature convergence,and the adaptive improvement can further improve the solving ability of genetic algorithm.