Research On Lot Splitting And Scheduling Method Of Multi-level Assembly Job Shop Based On Genetic Algorithm

Production scheduling is an important part in industrial activities.A reasonable production scheduling scheme can optimize the allocation of workshop resources,improve the utilization rate of equipment and reduce manufacturing cost,so as to improve enterprise competitiveness and market benefits.Most of the traditional researches focus on the job-shop scheduling problem JSP(Job-shop Scheduling Problem),while the assembly job-shop scheduling problem AJSP(Assembly Job-shop Scheduling Problem)as a more realistic extension of JSP,has received less attention.In AJSP,scheduling process is usually oriented to the product manufacturing with a multi-level tree structure,and the product completes the production process through two stages of processing and assembly.This kind of scheduling problem should consider not only the process sequence constraints of the tasks,but also the assembly association constraints among tasks at different levels.That is,only a complete set of associated parts can start assembly,so it has higher complexity and challenging.In this paper,this kind of problem with multi-level assembly structure is called the multi-level assembly job-shop scheduling problem,MAJSP(Multi-level Assembly Job-shop Scheduling Problem).Lot splitting as one of the important technical means in the production system,is widely used in the actual scheduling.Through reasonable lot splitting,the fluidity of sub-lot in the work can be improved,so as to shorten the completion time and improve the production efficiency.When lot splitting is considered in MAJSP,the sub-lot scheduling of parts at different levels will be more flexible and complex,because the random lot splitting of tasks and the assembly constraint relationship among the sub-lot tasks,it will bring great difficulties to the problem optimization and actual production management.Therefore,it is of great research significance and academic value to design and use efficient lot splitting strategies and scheduling algorithm to improve the coord ination among tasks and the flow of sublots in the multi-level assembly job-shop lot splitting and scheduling problem.First of all,this paper summarizes the research status of job-shop lot splitting and scheduling problem and assembly job-shop lot splitting and scheduling problem respectively.Based on the summary and analysis of the existing research deficiencies,this paper proposes the problem of lot splitting and scheduling problem for multi-level assembly products for the first time.Through the analysis of the difficult points in the the multi-level assembly job-shop lot splitting and scheduling problem,a mathematical model with the optimization goal of minimizing the completion time is constructed.Based on two kinds of solving strategies in lot splitting and scheduling: the step-by-step optimization strategy and the integrated optimization strategy,the overall solution flow framework of this paper is proposed.Then,according to the step-by-step optimization strategy,a hierarchical iterative optimization algorithm based on the combination of genetic algorithm to optimize the lot splitting and dispatching rules to optimize scheduling is proposed.The performance of different lot splitting strategies and dispatching rules in different product struc tures is compared through simulation experiments.Secondly,according to the integrated optimization strategy,the lot splitting scheme and the scheduling scheme are coded uniformly.The global field search algorithm and the feasible field search algorithm are proposed respectively based on genetic algorithm,in which the repairing operator is carried out in the global field search algorithm with the traditional genetic evolution operator to make the infeasible chromosome satisfy the assembly constraints be tween different levels and converted into the feasible chromosome.In the feasible region search algorithm,an improved population generation method,crossover operator and mutation operator are designed to make the genetic evolution process in the feasibl e domain.After that,the solution effects of three algorithms: the hierarchical iterative optimization algorithm,the global region search algorithm and the feasible region search algorithm,are compared through simulation experiments.Finally,a variety of lot splitting strategies are coupled to the feasible region search algorithm for solving.Among them,four lot splitting strategies are proposed to divide the workpieces into sublots in the two stages of processing and assembly,in order to explore the performance of lot splitting in different stages with different product structures.And according to their adaptability in different product structures,further put forward the collaborative complexity oriented strategy to guide the direction of lot splitt ing and to optimize the problem of large search space and low efficiency caused by the complete random search of lot splitting.