Research On The Method Of Product Modular Design For Supply Chain

Product design is anengineering based activity that realizes the customerrequirements into functions of a newproduct. Many studies have pointed out that theintegration of product and supply chainis a key factor for profitability and efficiency.Sinceit is the product design that determines which materials, components, and finishedproducts should flow through the supply chain, considering the available supply chainlocations, their capacity, costs and their demands while designing a product would helpdetermine the optimum product design and modular structure.This paper studies themethod of product modular design for supply chain, the purpose is to reduce the cost andlead time of supply chain while implementing product modularity.Based on the modular design and relevant theories, the difference between thegeneral process of modular design and the process of modular design for supply chainissummarized. The influence of supply chain cost and lead time in module divisionisanalyzed. The goal of modular design for the supply chainis clarified. The DesignStructure Matrix of parts’ comprehensive relationship is established by analyzing thefunction dependence and structural dependence of components.Amulti-objectiveoptimization model is constructed to maximize the product modularity,minimize the supply chain cost and lead time.Two methods based on Genetic Algorithmare used to solve the model and the mutation operator of genetic algorithm isimproved.The first method is using efficacy coefficient method to build thecomprehensive evaluation function of the three objectives. So the multi-objective problemis transformed into single objective problem, and then it is optimized by genetic algorithm.The second method isdirectly using the modified non-dominated sorting genetic algorithm(NSGA-II) to optimize the multi-objective problem.A refrigerator case is used to verify the method of the product modular design for thesupply chain. The Design Structure Matrix of parts’ comprehensive relationship isestablished by analyzing thedependence of refrigeratorcomponents.Theproductmodularization resultis respectively obtained by using two methods to solve the multi-objectiveoptimization model.The first method can get a best product modularizationscheme, and the maximum number of modules is analyzed by sensitivity analysis.Thesecond method can get a Pareto optimal solution set, which generate three productmodularization schemes，one of them is as the same as the result of the first method. Thevalidation and efficiency of this method are demonstratedby comparing with themodularization result of the traditional DSM method.