Research On The Job-shop Scheduling Optimization Methods And System Implementation For Complex Production Conditions

With increasingly keen market competition, each enterprise is searching for a good management system for production and operation to improve efficiency in production, operation and management as well, thereby enhance its own core competitive advantage. While the key of production and operation management is whether the production and scheduling process could achieve the optimal solution, the research on the production planning and scheduling is of great value both in thepry and in practice.There are two problems in the existing study of the JSP:First is regarding the processing parameters as fixed and precision values, and the JSP is considered as a purely mathematical problem. This is a strong assumption, which may cause severe difficulties in practice. In fact, there are many vaguely formulated relations and imprecisely quantified physical data values in real descriptions since presice details are simply not known in advance.Second is ignoring dynamic factors in the practical production process such as machine may break down, workers may get sick, deliveries may be delayed, new jibs arrive continually over time etc., making it necessary to make a scheduling even if only part of the problem is known, this is called dynamic scheduling. The paper focues on the compliex JSP with imprcision and dynamic environment, and develops an intellectual system which accordind to the production reality very well.The main works of the paper is as following:(1) The theory and methods of the compliex JSP optimization system. By reviewing the literature on optimization methods for JSP, the paper studies the optimization algorithms for the compliex JSP system, the new optimization algorithms-dynamic hybrid genetic algorithms (DHGA) based on the genetic algorithms and simulated annealing technique is proposed for solving compliex JSP. Computational results show that DHGA performs extremely well on both benchmarks and instances like industrial ones. The paper also discusses the influence of the parameters of GA using statistical analysis.(2)Construct the JSP model and the optimization methods in imprcision environment. The paper considers the fuzzy JSP with imprecise processing times. Signed distance and the index of optimism are used for constructing the fuzzy JSP models. The two fuzzy models are tested by solving benchmarks using dynamic hybrid GA, the results show the appropriateness of the two fuzzy JSP models. The uncertainty of the fuzzy makespan is considered and a bi-criteria optimization procedure for fuzzy JSP is presented. The uncertainty is measured by the spread of the triangular fuzzy number that represents the fuzzy makespan, the spread of the fuzzy makespan can be minimized in three different control ways and the practical situations of each control ways is defined.(3)Construct the JSP model theory and methods in dynamic environment.The paper considers the issue of robust and flexible solutions for JSP in dynamic enviroments. A robustness measure is defined and its properties are investigated. The concepet of rescheduling is described. Through experiments, it is shown that using GA it is possible to find robust and flexible schedules with a low makespan. These schedules are demonstrated to perform significantly better in rescheduling better after a breakdown than ordinary scheduling.(4)The paper develops a practically software E-proms which realizing production information management, the production plan arrangement automatically and the job shop scheduling control etc. The simple algorithms based on priority constraints and the DHGA are used to solving job scheduling in E-proms. The system has been used in Shenzhen Flextronics Company successfully.