本文嘗試在有限空間下建構一多目標生產系統設計模型 (Multi-Objective Production System Design，簡稱MOPSD模式)，本模式不僅可決定整個生產系統單位時間的總物料運輸流量最小化之佈置且可追求考量同步生產情況下之最適系統生產周程時間之最少工作站指派數。MOPSD模式之第一目標為一二次分配問題(Quadratic Assignment Problem, QAP)，在傳統有關生產單元佈置之QAP問題中常僅考慮單一生產線之佈置，有關生產單元的規模、多生產線以及有限空間佈置之問題鮮少被提及。因此本研究的第一階段模式利用空間離散處理以使空間中的可佈置位置有限化，以求得總物料運輸流量最小化之二維佈置。再者，由於傳統的ALBP (Assembly Line Balancing Problem)之探討，鮮有考量同步生產的情況，因此會造成生產率被低估的情形產生。所以本研究第二目標將同步生產的情況加以考量以避免造成生產率的低估，進而改善資源及設備浪費的情況。此外，本研究藉由四位置編碼技術(Four Position Coding Technique)將第一階段與第二階段模式成功地連結並使用同步生產的考量加以落實。最後本研究以一數值範例，仔細地說明 MOPSD之求解步驟。綜言之，MOPSD因採用Lingo9.0 extended version之方法建構模式，而使本模式具有高度的重現性(Repetitivity)，各工廠可根據其不同的情況，利用本模式，求取最佳解。因此本研究對實務上多生產線生產系統佈置提供了一項具價值性的決策工具。 This paper proposed a mathematical model called Multi-Objective Production System Design (MOPSD) model and a two-staged solving procedure. The MOPSD model not only considers the two-dimension layout of a production system under the constrained factory space to achieve the minimal total material transportation flow but determines the optimal assigned number of workstations and the optimal system cycle time by concerning the simultaneous production. The major objective of MOPSD model is regarded as a Quadratic Assignment Problem (QAP), but it considers the size of operation element, the multiple production lines, and the constrained two-dimensional factory space where traditional QAP studies are seldom mentioned. The second objective of the MOPSD model is heading to determine the optimal system cycle time as well as the optimal assigned number of workstations under the simultaneous production situation for preventing the underestimation of the system production rate.The first-stage solving procedure of MOPSD model applies the space discrete technique to make the infinite deployable locations become finite, and then we can achieve the major goal through the system layout. In addition, the Four Position Coding Technique (FPCT) is developed to code the nodes in the deployed production system obtaining from the first stage, and therefore those FPCT codes can function as the input parameters in the second-stage solving procedure. These two solving procedures are constructed by the syntax of Lingo 9.0 extended version, and a numerical example is followed to describe the detailed solving procedure. In summary, the repeated characteristic of MOPSD model and its solving procedure make this study be a valuable decision support tool because they are constructed by the packaged software Lingo 9.0. That is, a factory can easily duplicate to solve other cases by changing its input parameters only.
