Research Of Multi-objective Flexible Job Shop Dynamic Robust Scheduling

The research of job shop scheduling and optimization technology has become the foundation of advanced manufacturing technology. In the manufacturing workshop, the scheduling problem is very complex. Job shop Scheduling problem is usually a multi-objective optimization problem, and there are often conflicts between different goals. In addition, there are uncertainty disturbance factors in the process of actual production, such as machine fault, changing process time, rush orders and so on. Therefore, deeply research of job shop scheduling problem, can better guide the production.Based on this background, this paper studied the dynamic scheduling problem of multi-objective flexible job shop, and obtained some meaningful results.The main work of this paper is as follows:(1)This paper summarized the background, research status at home and abroad, the trend of job shop scheduling problem; analyzed the existing workshop scheduling algorithms; elaborated the research significance and research purpose of this subject.(2)In this paper multi-objective optimization algorithms were analyzed, and compared evolutionary algorithm with the traditional algorithms, emphasized the advantage of evolutionary algorithm. Based on the different targets, an evaluation index system of multi-objective flexible job shop is put forward. The system contains several flexible job shop scheduling targets, such as time, machine load, cost and delivery time. And the calculation methods of each target were discussed.(3)On the target of minimizing maximum completion time and minimum advance/tardiness penalty, the multi-objective flexible job-shop scheduling model is established. In this paper, a multistage dynamic robust scheduling policy which includes evaluation of disturbance event, buffer integrated, local update, global scheduling was put forward. This strategy can effectively reduce the number of global scheduling, guarantee the continuity and robustness of the scheduling scheme.(4)Genetic algorithm of multi-objective flexible job-shop scheduling problem was improved in this paper. Immune algorithm was combined with genetic algorithm, using the immune system and entropy principle to maintain the diversity of population. In addition, in view of the lack of elite selection strategy of multi-objective genetic algorithm, a distribution function was import. At last, the feasibility of this algorithm was verified by an experiment.(5)According to the characteristics of the actual manufacturing plant, this paper proposed a multi-objective immune genetic algorithm based on rolling windows strategy. Based on periodic and event-driven scheduling strategy, dynamic process was divided into a series of continuous static scheduling interval, in each interval using multi-objective immune genetic algorithm which based on Pareto optimization strategy. Based on goals in scheduling model, job selection principle of the artifacts window was put forward.(6)According to the characteristic of flexible job shop, an extension deviation index was designed. The index contains job index and machine index, and it works together with multistage dynamic robust scheduling algorithm.