Research On Algorithm And Simulation Of Heuristic Dispatching Rules To Job Shop Scheduling Problem

With the increasing sophistication of FMS and CIMS, job shop scheduling problem (JSP) has necessitated an improvement in the way companies manufacture their products. According to statistics,it is approximately 90% enterprise belonging to make-to-order enterprise in our nation manufacture industry. However, in the make-in-order manufacture conditions, the "one-time" and experiential property of production lead to a deviation of the plan in job shop level, or even more frequent conflict of changes and accidents. The management on the job shop level has become the weakest part in the management of manufacture system, as well as the worst problem a supervisor in the manufacture company may face. Therefore, the dynamic job shop scheduling problem in a typical manufacture system is major analyzed in this paper. Details are shown as followed:First, this paper present efficient dispatching rules, their classification and model are induced, followed by heuristic algorithm of job shop scheduling problem were proposed. It is quite common to encounter a dynamic and stochastic arrival of jobs in manufacturing systems and hence JSP is a dynamic scheduling problem. It is important for a dynamic scheduling problem to the rapidity of scheduling and the agility of response. The current state of dynamic scheduling problem always can not be solved by exact optimization algorithms. Therefore, in many real-life situations, dispatching rule is easy to implement and low computation complexity, so dispatching rule is always used to solve JSP.Second, based on dispatching rule we focus on researching and designing heuristic algorithm. And simulation scheduling platform is designed by object-oriented programming language Java based on JBuilder software platform.Third, this paper introduce to simulation experiment design and the analysis of simulation scheduling conditions. A review of the different studies is the performance objectives in the relative of scheduling literature. The results of scheduling in 16 different simulation experimental conditions are gain.Final, we carry out the statistical analysis of the scheduling results. We presents analysis of 7 different kinds of performance objectives for evaluating 11 dispatching rules, such as maximum flowtime,mean flowtime,variance of flowtime,the percentage of tardy jobs,the number of tardy jobs,maximum tardiness and mean tardiness. The different performances of different dispatching rules are sum in different simulation experimental conditions, which provide the valueable guiding significance for heuristic rules applied to a dynamic job shop scheduling problem.