在風險偏好考量下的改良群體PROMETHEE模型之建構

機構典藏 > College of Business and Management > Department of Management Sciences > Thesis > Item 987654321/105318

Please use this identifier to cite or link to this item: https://tkuir.lib.tku.edu.tw/dspace/handle/987654321/105318

Title:	在風險偏好考量下的改良群體PROMETHEE模型之建構
Other Titles:	A study on the improved group PROMETHEE model based on DMs’ risk preferences
Authors:	張瑜庭;Chang, Yu-Ting
Contributors:	淡江大學管理科學學系碩士班時序時;Shih, Hsu-Shih
Keywords:	PROMETHEE;S-shaped價值函數;群體決策;模擬;多準則決策分析;S-shaped value function;group decision-making;simulation;MCDM
Date:	2015
Issue Date:	2016-01-22 14:53:06 (UTC+8)
Abstract:	本研究以展望理論中的價值函數整合至多準則決策分析方法中，並以偏好順序結構評估法(Preference Ranking Organization METHods for Enrichment Evaluations；PROMEHTEE)，發展具有風險偏好考量的群體PROMEHTEE模型。同時以台灣電子廢棄物回收處理廠為研究案例，進行績效評估。本研究擴充原PROMETHEE偏好函數模型範圍為[-1, 1]，使偏好函數具有正負向值考量，再轉換成S-shaped價值函數估算其損失與利得，納入PROMETHEE運算步驟中整合。並使用群體分析，以了解群體決策過程如何綜合意見、衡量矛盾點及其改善。為了解模型之穩定性，使用敏感度分析以測試各項門檻值對結果之影響，及運用統計等級相關檢定，以比較其排序結果之相關性，同時使用電腦模擬分析以試驗績效值對結果的影響。結果可發現加入風險偏好價值函數後，其淨流量值擴大，雖不影響最適解排序，但使中間方案排序分辨性提高。其中於統計檢定可發現基於傳統方法下之改良其結果可有效表現決策者偏好；於敏感度分析可知，當門檻值與價值函數參數異動時，其結果異動不大，代表修改模型具有穩定性。於電腦模擬之結果可發現，透過風險價值函數轉換後為最佳方案之機率較為分散，其機會成本部分可能為決策者所低估，而分群結果與原始報告內容稍有所不同。 This study proposes an improvement for Preference Ranking Organization Methods for Enrichment Evaluations (PROMEHTEE) based on prospect theory of the S-Shaped value function to construct risk preferences in single decision-maker (DM) and group PROMETHEE model. An example of evaluating E-waste recycling plants in Taiwan is illustrated. We extend the range of preference value function within [-1, 1], and use S-shaped value function to evaluate loss and gain to bring into integration rank of PROMETHEE method. In order to evaluate for handle decision-maker’s risk preferences, and simulation group PROMETHEE method to understand the procedure of group decision. We not only use PROMETHEE III to analysis, but also apply the model testing, such as sensitivity analysis is to test the threshold and parameter that inference to results, ranking test is for comparing the ranking results in correlation, and computer simulation analysis is for testing the influence of value functions. The result shows that the net flows are extended when we integrate the value function of risk preference. Although the preferred solution is the same, the middle ranks are enhancing the distinguishing. The statistic test can be found that the modified PROMETHEE is useful performed the DMs’ risk preferences based on traditional PROMETHEE, the sensitivity analysis can be found that the results does not change too much, it represents the stability of the modified model and can express the preference of DMs. Through the value function not only found that probability of preferred solution is more decentralization, but also can know that opportunity cost may be elided by DMs. Finally, the clustering of simulation are different from EPA’s reports.
Appears in Collections:	[Department of Management Sciences] Thesis

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	137	View/Open

Loading...