A Novel Dynamic Center Particle Swarm Optimization Algorithm to the R&D Multi-project Scheduling Problems in Mass Customization Enterprise
In the fierce-competing 21st century, the primary factor of winning marketplace and customer is to enhance the R&D ability of enterprise, so that enterprise can satisfy varying customization need of customer with the lowest cost and shortest time span. In order to achieve this purpose, enterprise must allocate resource effectively to the R&D Multi-Project and schedule activities properly. Based on the analysis of features of R&D Multi-project environment and resource allocation demand in MC (mass customization) enterprise, this paper improves the multi-project scheduling model proposed by Wiest and Levy（1977）and develops a R&D multi-project scheduling model with crashing in MC environment based on delay penalty. Meanwhile, based on the analysis of PSO (particle swarm optimization) algorithm and its improvement by other authors, we proposes a novel dynamic center particle swarm optimization algorithm in which inertia weight and learning factors:c1,c2 in particle velocity update processing vary along with the increase of iteration.w varies at random in the range of 0.4,0.9. But c1,c2 vary linearly along with iteration. That is, large particle c1 refers to social learning at first. With iterating,c2 increases and c1 decreases. Particle refers to cognitive learning to make a scrutiny in local scope. This algorithm is used to solve multi-project scheduling model. Then we simulate the variant example in reference 3 through mat lab program and obtain optimal solution, so it proves the feasibility of model and algorithm proposed in this paper.
R&D Multi-project Project Scheduling PSO Dynamic Center PSO
WU Juan SHAN Miyuan PENG Danni
Business and Management College, Hunan University. Changsha, Hunan, China
国际会议
The Fifth InternationalSymposium on Management of Technology(ISMOT07)(第五届技术与创新管理国际研讨会)
杭州
英文
877-881
2007-06-01（万方平台首次上网日期，不代表论文的发表时间）