Application of soft computing to the merchandise management

• Main Authors: Chien-Yuan Lai, 賴鍵元
• Other Authors: 張百棧
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/25267627619235006193

博士 === 元智大學 === 工業工程與管理學系 === 94 === Merchandise management plays a crucial role for successful enterprises. Forecast becomes critical when strengthening the functions of merchandise management. As forecasting has in itself uncertain factors like complexity and being unstructured, the present thesis proposed related Soft Computing techniques to provide managers with aid in the applications to decision making. In order to demonstrate that the methods proposed in this research stand out in forecast capability, book wholesalers are used as a case, with their sales of new products and returns of slow-selling goods as two major subjects in merchandise management. Presented, respectively, to deal with these two subjects are the forecasting method for sales of new released books combining Self-organizing Maps (SOM) with Case-based Reasoning (CBR) method, and the forecasting method for returns of slow-selling books blending the Genetic Algorithm (GA) with the CBR method. The experimental results demonstrated that these hybrid methods this research proposed outperformed other methods in various forecasting in terms of accuracies. With the achievements of this research summed up, it is known that the soft computing will have notable potentials to be applied to problems in the merchandise management.