SPAN群體決策方法的改良及其在公路系統改善問題之應用

Abstract

SPAN群體決策法改善了傳統直接評分法忽略個人判斷能力差異性之問題，及無法凸顯專業者意見之缺點。 然在資訊不完全之情況下，以SPAN Method做決策將可能產生錯誤判斷或資訊未被充份利用的情形。本研究的動機乃在改良SPAN Method以適應資訊不完全之情況，研究目的則在於提出一資訊不完整下的群體下的群體決策模式。在此模式中，各決策者只對熟悉的方案評定權重，對不熟悉的方案則指定參考對象，然後模式將此決策問題轉化為一數學規劃問題，進而求其全域的最佳解。 本研究最後擬以公路系統改善為例，說明本研究所提的群體決策模式之運作。

The SPAN Method has improved the traditional direct voting method that ignores the different abilities of judgment between indiviuals and can not reflect the opinions of experts. However, when the information is incomplete, using SPAN method to make decision could lead to the wrong judgment or insufficient employment of collected information. The motive of this research is to improve the SPAN method to hit the circumstances of incompleteness of the information. In this paper, we present a new group decision model in which each decision maker is only asked to give weights to the alternatives familiar to him and to distribute the quota to the members he will cite for the unfamiliar ones. Then this decision making problem will be transferred to a mathematical programming problem, and then to find its global optimum. We take the example of improving highway system as our example to explain the operation of our model.