頁籤選單縮合
題 名 | 基因演算法在配置陳列空間上的應用=The Application of Genetic Algorithms to Shelf-Space Allocation |
---|---|
作 者 | 楊銘賢; | 書刊名 | 管理學報 |
卷 期 | 16:2 1999.06[民88.06] |
頁 次 | 頁231-253 |
分類號 | 496.6 |
關鍵詞 | 基因演算法; 陳列空間配置; 資源規劃; Genetic algorithms; GA; Shelf-space allocation; Resources planning; |
語 文 | 中文(Chinese) |
中文摘要 | 如何有效配置陳列空間是零售管理的重要課題,本研究對此離散最適化問題提出 一個一般化數學型式及一個簡化的整數規劃型式,並發展出求解的基因演算法,分別是採不 同再生方式的 GA-1, GA-2 GA-3。本研究以模擬問題來驗證這三種方法的有效性,分別 用基因演算法求得解的利潤與CPU 時間對最適解的比值作為績效指標,結果顯示三者的績效 都令人滿意。整體而言,以採菁英再生之二階段再生方式的 GA-2 績效最佳,若取重複求 解十次之最佳為解,其在二種型式問題的平均利潤比分別高達0.989與0.967。本研究的求解 方法除可用來配置陳列空間外,亦可用於有關資源規劃的其他管理決策問題。 |
英文摘要 | Allocating shelf-space effectively is very important to retail management. This study presents a general model and another simplified integer programming omdel for this combinatorial optimization problem. Genetic algorithems GA-1, GA-2, and GA-3 that use different operators of reproduction are also proposed to solve this problem. The ratios of profits and CPU times of their solutions to the optimal solutions are used as the measures for performance validation. The results of simulation tests show that hteir performances are satisfactory. The performance of GA-2, which uses the two-stage reproduction of elitist reproduction, is the best. If the best among ten replications are chosen to be the solution, its average profit ratio will be 0.989 and 0.967 for two different models respectively. In addition to the application to shelfspace allocation problems, these methods can also be applied in other management decisions related to resources planning. |
本系統中英文摘要資訊取自各篇刊載內容。