Research On Scheduling Of Whole-set Orders In JSP Based On Differential Evolution Algorithm

As one of the most difficult combinatorial optimization problems, Job-shopScheduling Problem has always been a hot issue. Researching and applica tion of effectivescheduling algorithms and optimization technology, ma nufacturers can improve efficiencyof production and lower costs. Differential Evolution algorithm, as a new intelligentalgorithm, has a great ability of find ing globa l convergence robustness and no need of thecharacteristic information. Focusing on the differentia l evolution algorithm, this thesis hasmade an in-depth research on the Job-shop Scheduling Problem with whole-set orders asthe objective.Firstly this thesis introduces the research studies and related algorithms of job-shopscheduling problem, as well as problem of whole set orders. Then, a set of mathematica lmodels is established. To speed up the convergence of the algorithm, a hybrid differentia levolution algorithm based on a neighborhood search is proposed. Contrast with standardDifferential Evolution Algorithm, results of severa l simulation experiments show that ahybrid differentia l evolution algorithm based on a neighborhood search has a great abilityof find ing optimization solution and faster ability of convergence. To solve the problem ofma ximizing the weighted whole -set orders and minimizing the weighted job which aredela yed, a mathematica l model is establish. An improved differentia l evolution algorithmbased on dual-population is proposed. Contrast with standard Differential EvolutionAlgorithm, results of severa l simulation experiments show that the improved differentia levolution algorithm based on dual-population has a great ability of find ing optimizationsolution, and has faster ability of convergence.