結合邏輯斯迴歸與灰關聯於易腐性商品訂購策略之研究

對企業而言，最佳訂購與存貨政策是降低企業營運成本最有效且最直接的方法。如何求得正確的訂購量，尤其是需求變動劇烈的產品，一直以來都是企業所重視的一環。本研究以統一超商的便當為分析對象，針對便當商品發展出一套有別於一般依賴POS系統決定訂購量的三階段訂貨策略。首先，本研究以報童模式求得粗估的各類便當數量，並進而決定整體便當之基本訂購量。其次，運用邏輯斯迴歸分析，考量各種可能影響需求之因素，對便當整體基本訂購量提出微調總量的建議。最後，利用灰關聯分析法分析各式便當之獲利績效，作為各式便當訂購數量增減之依據，依此決定各式便當最終訂購量。本研究係運用邏輯斯迴歸二元判別之特性，提出具創意的想法，用於判定訂購量會報廢或不會報廢，並藉此探究所有可能影響訂購量之因素。經邏輯斯迴歸分析可知，天氣與來客數是影響需求變異的要因。由於本研究係植基於存貨管理中的風險共擔概念，先將整體便當的需求彙總，以降低需求變異性，並提高預測準確度。續而使用績效分析進行資源分配，再求出各式便當的需求，故可使利潤極大化。此外，本研究將許多可能影響銷售量的因素納入考慮，故更能掌握需求的變異。此套訂購決策流程不僅可運用到具隨機性需求的易腐商品，為超商管理者帶來更高的獲利水準，對於掌握其他需求變異大的產品同樣可以發揮很好的效果。

關鍵字

易腐性商品；報童模式；邏輯斯迴歸；灰關聯分析；訂購策略

並列摘要

The key to the success of a convenience store is the ability to make decisions that not only follow consumer needs, but also reduces operational costs such as efficiently control their stock replenishment, especially for perishable items such as meal-boxes. To solve this problem, we proposed an innovative decision support system to determine the optimal amount of replenishment. In the first step, we obtained the basic order quantity of the overall meal-box by newsboy model. The basic order quantity may not actually match with the real demand due to the effect of uncertain factors such as the climate and promotion activity of substitute products. Therefore, in the second step, a novel warning system is established by employing the logistic regression model to modify the basic order quantity. In the third step, we employ GRA to allocate the optimal order quantity of each kind of meal-box. Using actual data from a convenience store which is a part of the President Chain Store Corporation in Taiwan, the prediction accuracy of the decision support system was evaluated. Through numerical experiments, that the proposed policy can accurately determine the optimal order quantity is confirmed.