使用共同權重於多指標績效評量

Abstract

過去有許多的文獻探討多目標的績效考核方面的研究。此篇研究考慮三個績效評估的情境，在每個情況下，我們發展了一系列的程序，使得單位組織可以依據一套共同的評量標準，獲得彙總分數來做排名。在第一種情境中，單位績效評價具有多個投入和產出指標，使用五階段的資料包落分析法(Cooper et al., 2006)的數學規劃模型來求解。第二種情境，為提高整體組織之生產力，在考慮實務應用上的限制下，重新分配各子單位的投入資源與產出目標。在第三種情境，以奧林匹克為例，用各國家所獲得的金、銀、銅牌，以模式獲得的各獎牌權重來計算這些國家總成績，藉此發展投票系統中權重求取模式。此外，延伸發展新的多準則投票系統來解決群體決策問題。

There is abundant literature on performance evaluation for multiple units with multiple criteria. Three scenarios of performance evaluation are considered in this research. In each scenario, we developed a procedure to determine a common set of weights to the criteria so that the units could be ranked according to their aggregated scores. In the first scenario, units are evaluated with multiple input and output indices. The procedure consists of the five phases of mathematical programming models of data envelopment analysis (Cooper et al., 2006). The second scenario considers the reallocation of inputs and adjustment of outputs with side constraints so that the entire units’ productivity could be improved. In the third scenario, we use the number of gold, silver, and bronze medals that nations won in an Olympic Game to deliver the method that determines the three weights for the medals such that those nations could be ranked according to their aggregate scores. We also develop a new multiple criteria voting system in group decision problems.