在設施佈置問題中,目標因子依其特性,可區分為定量因子及定性因子。定量因子,如距離、流量、成本等可量化數值之因素;而定性因子,則如部門間之相關性及環境效應等。過去的研究趨勢多以定量因子所組成之單一目標及多目標為主要研究方向,但目前產業逐漸走向服務業,其空間組成多以辦公室、服務區為所需區塊,其特質不如製造系統中,有可量化之流量、成本數據可直接套用。本研究以部門間之相關性之單一目標為主要方向,並考慮距離因素,嘗試提出一改良的目標模式,並應用啟發式演算法之 ”蟻群尋優法” 與應用空間填滿曲線之特性,發展一設施佈置軟體,找出一合理佈置方案。其改良後的目標式與以往TCR及WCR評估方法比較,結果顯示佈置方案以所得評分排名,TCR數值僅能大致區分出數個族群,而WCR與本研究之目標值,較可細分別其優劣。另外,本研究應用蟻群尋優法求解設施佈置問題,經由田口實驗的方式,求取案例的最佳參數組合,期望能增加其求解效率,求取一較佳解。其求解品質會隨著部門增加而降低,建議可增加世代數及螞蟻數來提升。
In the facility layout problems, the objective functions can be classified into either quantitative or qualitative factors in nature. The quantitative factors include distance, flow, or cost. However, the quantitative factors may include the departmental relationships, or environmental effects. In the past, the researches tend to focus on the quantitative factors with single or multiple objectives. With the rapid growth of the service industry, the focus of the facility layout problem is concerned with the best allocation of the offices and service areas. However, this problem does not have the flow information and should take the departmental relationship instead. This research tried to find an appropriate model which focused on the departmental relationship with single objective. An Ant Colony Optimization (ACO) heuristics was developed to find the best layout alternative with the aid of the space filling curve in the departmental generation and allocation. The results showed that this performance rating of the proposed heuristic (Z) performed evenly with the TCR rating, while the performance rating (Z) would perform better with the WCR rating. We also applied the Taguchi method to find the best design parameters of the ACO heuristic so that a better solution can be found. The solution quality was affected by the number of departments, and can be improved by increasing either the generation or the number of the ants.