Solving stochastic production planning by using decision tree analysis

• Main Author: 麥姿穎
• Other Authors: 洪一峰
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/5a7b5x

碩士 === 國立清華大學 === 工業工程與工程管理學系 === 92 === Production planning problems have many uncertain factors. Most of production planning models use deterministic parameters to simplify production plan problem and in order to find production plan quickly. This study considers uncertainties in production planning problem. Thompson and Wayne (1990) proposed an integrated modeling approach which can only provide the expected value with perfect information (EVwPI), but it cannot provide a production plan. This study applies decision tree analysis to solve stochastic production planning problem. We express production planning problem with stochastic demand. After a decision tree being constructed, we can compute the expected objective value by using backward induction procedure to find an optimal production plan. When a production plan must be calculated, we can input new product demand forecast information, and reconstruct the decision tree to find the production quantity of the next period. The result of our experiment shows that the decision tree analysis not only be able to find a set of production plan quickly but also make the expected value of the decision tree analysis close to the expected value with perfect information.