Research On Differential Evolution Method For Facility Layout Of Workshop

With the development of market demand and the enterprise production, the traditional facility layout of workshop and backward production form have become the severe barriers to the development of our manufacturing enterprises.Recently, the meta-heuristic algorithms have been becoming an important research field and applied to solve the complex engineering problems extensively. Based on the facility layout of workshop, this thesis makes a research on Differential Evolution method and application. The method is used to solve two kinds of layout problems, One of which is the linear layout of workshop problem, the other is the loop layout of workshop problem. Both of them are the basic units of complicated facility layout of workshop, which have certain significance.This thesis adapted from National NaturalScience Foundation of China (grant nos. 60674078 and 50575031) and the National High-Tech R&D Program of China (grant no.2006AA04Z109),which taking the facility layout of workshop in Dalian SANYO Compressor Co.Ltd.Thus, a Differential Evolution method with Random Key is presented.(1) This thesis presents a kind of Differential Evolution Algorithm with Random Key method(RKRDE) for solving linear and loop facility layout problems of workshop. The algorithm transforms continuous optimization problems to discrete optimization problems.At first,it uses real-coded for crossover and mutation.Secondly,it decodes before solving evaluation function.Then it sorts by serial number of facilities. At the end,it selects a optimum scheme.By this means,it speeds up convergence and improves result quality.(2)This thesis realizes the algorithms including RODE, Differential Evolution Algorithm with sub-range encoding(SRRDE), Simulated Annealing Algorithm(SA), Tabu Search Algorithm(TS), Genetic Algorithm(GA). The model of the linear and loop facility layout samples of workshop is builded model by the simulation software 'Witness'.The experimental results comparing with five algorithms show that RKRDE is effective to solve facility layout problem.It's hoped that this study would be beneficial to the theoretical research of the Differential Evolution Algorithm. Meanwhile, it is also expected that the proposed approach (RKRDE) would be expanded into the applications of other complex engineering design problems such as the facility layout problems of workshop.