Application Of Improved PSO Algorithms To Job-shop Scheduling System

Scheduling theory belongs to combination optimization problem. Introducing the optimization theory to the field of job scheduling in workshop can improve the performance of algorithm, make the algorithm apply to broader fields and complete the whole system of the algorithm, which is a subject including both theoretical meaning and practical values.The parallel genetic algorithm has always been the hot spot of the research on genetic algorithm. This article proposes an improved parallel genetic algorithm based upon the traditional coarse grain genetic algorithm. The proposed algorithm improved the traditional one mainly in two aspects: on one hand, it employs a mechanism which can dynamically adjust each island's migration rate according to its average fitness. Therefore, if one island's average fitness is high, its diversify will be increased by increasing other islands'migration rate to this island in order to avoid prematurity; the improved algorithm also calculates the surviving period for each chromosome involved in migration, not only enable the chromosome with high fitness to survive longer but also avoid the explosion of the population of some island. These improved measures are of great significance on avoiding prematurity of the algorithm, as well as increasing global searching ability of the algorithm. The improved algorithm is tested through using examples in the standard test set and the actual problem in scheduling. The simulation results of the test show that this algorithm can complete the task of the fast search in the given range and global optimization.In the meantime, according to the practical situation of the manufacturing business, this article designs a new method to initialize the population which can apply to practical job shop problem. If compared with the conventional population initializing algorithm which only randomly divides the initial population to many smaller groups, the new algorithm is much more effective in the aspect of increasing the diversity of the population.This article designed and implemented an intelligent scheduling knowledge base system platform for workshop, which is targeting at some factory's practical problem. Furthermore, the improved algorithm has been embedded into the knowledge base system. If we use this new method to solve the practical problem, the result shows the solution is both feasible and effective。...