請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/50722
標題: | 多群體與多變異下的偏好演化 On the Evolution of Preferences: Multi-population and Multi-mutation Scenarios |
作者: | Yu-Sung Tu 凃又菘 |
指導教授: | 莊委桐 |
關鍵字: | 偏好演化,演化穩定性,非對稱賽局,可觀測性,多重突變,端點,非合作賽局報酬區域, evolution of preferences,evolutionary stability,asymmetric game,observability,multiple mutations,extreme point,noncooperative payoff region, |
出版年 : | 2016 |
學位: | 博士 |
摘要: | 我們使用間接演化方法來研究偏好演化。假設每位參與者對所有可能的結果皆有個人的偏好,並且他們依此偏好及所獲知的對手資訊來做最適選擇。不過他們實際所獲得的報酬卻是被一個未知的償付機制所決定。
在第 2 章,我們分別根據不同的資訊透明度來探討多群體環境裡的偏好演化及其所衍生的結果。假設每位參與者有 p 的機率可觀察到對手的資訊。我們分別找出當 p = 1 和 p = 0 時,偏好演化穩定的必要及充分條件。我們也檢查了當兩 p 值被擾動後,上述結果在純策略的情況下是否保持不變。 在第 3 章,假設對手資訊皆可以被觀察。我們分別在多群體及單一群體環境裡探討由多重突變所引起的偏好演化問題。同時發生多個偏好變異可視為篩選演化穩定偏好的一種方式。我們分別在這兩個群體環境設定下找出多重變異偏好演化穩定的必要及充分條件;我們分別描繪了演化穩定等級與效率階層間的關係。 在第 4 章,賽局報酬區域的一般性質將被進一步研究,尤其是非合作賽局報酬區域。我們藉著在非凸集合上定義其端點這一新的方法來分析非合作賽局報酬區域;我們利用其端點及支撐超平面來刻畫此區域。根據此章主要的定理,合作與非合作報酬區域的端點皆可用純策略報酬組合來表示,並且這性質可幫助我們得到許多有意義的結果。 We use the indirect evolutionary approach to study the evolution of preferences. Each player has subjective preferences concerning outcomes, and chooses a best response based on his own preferences and the information about the opponents. However, the players' actual fitness for each strategy profile is assigned by a material payoff function. In Chapter 2, we study the features of stable outcomes in a multi-population setting under various observability. Players may observe their opponents' preferences with some fixed probability p. We examine necessary and sufficient conditions for evolutionary stability for p = 1 and p = 0. We also check the robustness of these results against a small perturbation in p for the case of pure-strategy outcomes. In Chapter 3, we study the evolution of preferences under complete information with multiple mutations in multi- and single-population settings, respectively. Multiple mutations would be introduced to refine the evolutionarily stable outcomes. In the two population settings, respectively, we examine necessary and sufficient conditions for stability against multiple mutations; we characterize the relations between the order of stability and the level of efficiency. In Chapter 4, the range of a payoff function in a game is investigated, especially the noncooperative payoff region. A novel tool, the notion of extreme points of non-convex sets, is introduced to analyze noncooperative payoff regions; we try to characterize the noncooperative payoff region in terms of its extreme points and supporting hyperplanes. The main theorem says that all extreme points of a payoff region can be generated by pure-strategy profiles, and it helps us to achieve meaningful results. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/50722 |
DOI: | 10.6342/NTU201600975 |
全文授權: | 有償授權 |
顯示於系所單位: | 經濟學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-105-1.pdf 目前未授權公開取用 | 923.79 kB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。