| The process of semiconductor manufacturing is highly complex and capital intensive. The complexity in semiconductor industry is mainly due to multiple product mix, complicated manufacturing process, and requirement for high machine utilization. In addition, In the semiconductor industry, the determination of the number of manufacturing tools needed to manufacture forecasted product demands, is particularly difficult because of its sensitivity to product mix, the uncertainty in future demand, the long lead time for obtaining tools and large tool costs. Therefore, production planning and capacity planning of semiconductor manufacturing system is a very complicated problem.The problem of production planning and capacity planning in semiconductor manufacturing system is a typical multi-objective problem with characteristics of multi-restriction, plenty of variables and uncertainty. The existing optimization method is deficient both in theoretical analysis and practical application, which need to be studied deeply and systematically. Based on the research and analysis of Multi-Objective Optimization theory (MOO), Multi-Objective Evolutionary Algorithm (MOEA) and Particle Swarm Optimization (PSO), this paper proposed a strategy of applying PSO to solve multi-objective optimization problem. The MOO model of semiconductor production planning is proposed considering minimizing unmet demand and maximizing utilization of capacity and the solution strategy of two-layer PSO is developed to solve the problem. We formulated scenario based multi-objective stochastic programming model to describe the problem of capacity planning under uncertainty and applied improved Multi-Objective PSO (MOPSO) to solve it. Through the example simulation, the result indicated that MOPSO in this paper works well to solve the problem of production planning and capacity planning of semiconductor. |
PDF Full Text Request