透過您的圖書館登入
IP:18.188.142.146
  • 學位論文

專家系統在批量問題上的應用

An Application of Expert System on Lot-Sizing Models

指導教授 : 葉若春 黃崇冀
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


對於單階層、單物項之訂購批量模式而言,除了由純數理推導所得,一直 不斷地有學者提出啟發式批量模式,這些啟發式解法雖不能保證得到最佳 的規劃,但求解過程卻簡單易懂,易於為企業所了解與使用。本文針對數 種啟發式批量法,討論其在不同的生產形態下的成本效應。為此,本文設 計了一實驗,以下列三種因素: 1.需求的變化情形。 2.存貨成本。 3.訂貨成本。來模擬不同的生產形態, 觀察在特定的生產形態下 LFL 、 EOQ 、 POQ 、 LTC 、 LUC 五個批量模式的表現。但現實的生產形 態可能有千百萬種,以實驗中有限的實驗值將難以滿足,若要勉強應用, 可能會發生偏差。為解決此問題,本研究採用專家系統中不確定推理的確 定因子 ( Certainty Factor )與內分點公式,以突破實驗值有限的限制 ,使得不在實驗值預定的生產型態,亦能運用實驗結果.

關鍵字

批量 存貨 不確定性 確定因子 專家系統

並列摘要


Scholars continuously proposed heuristic lot-sizing models except pure mathematical ones for the lot-sizing models of the single level and single item.These heuristic lot-sizing models could not be assured to get the optimal solutions, but their processes of solving problems are easy to understand and accept for the enterprise. This research is intended to study the performance of lot-sizing models in different manufacturing environments. I designs an experiment with following three factors to simulate in different environments: 1.demand 2.carrying cost 3.ordering/setup cost The experiment tests five lot-sizing models:LFL 、 EOQ 、 POQ 、 LTC and LUC in different manufacturing environments. There are numerous environment types in the pratical situation.The experiment could hardly be fit completely.To solve this problem, this research applies Certanty Factor in uncertainty inference of expert system to break the limits of the finite production types. By that, we can still use the result of experiment to get the performance of lot- sizing methods even if the production type is not in the experiment.

延伸閱讀