朝向整合演算式賽局理論及學習與社群網路

Abstract

我先前在攻讀電腦科學博士期間的研究主要在分析由各種不同個別策略行為(strategic behavior)和均衡概 念(equilibrium concepts)造成的整體效率變化(系統效能下降), 更明確地說, 是在賽局理論環境下利他(altruism) 和害他(spite)行為和各種不同均衡概念的影響◦ 我們已發展和分析了一個在一個社會或經濟/社群網路環境中具有利 他和害他參與者的賽局理論模型◦ 中長期來說, 我計畫逐漸發展 “實際世界中”的演算式賽局理論和機制設計, 目前 包括除利他和害他之外其他從社會的角度探討的議題像是公平(fairness)和互惠(reciprocity), 以及重覆賽局中的學 習(learning in repeated games)問題◦ 廣泛地說, 我想要使用演算/數學的分析技術和計算/實驗的方法處理經濟和 社會的問題◦ 在這一年的提案中, 我先總結先前的結果, 伴隨著一些可以在前半年進一步研究的問題(第一部份), 然後再提 出更多可以在前半年初探並在後半年完全探討的問題或方向(第二部份)◦

My previous focus during PhD study in Computer Science is analyzing the global inefficiency (performance downgrade) caused by various individual strategic behaviors and equilibrium concepts, specifically, the effects of altruism and spite as well as various equilibrium concepts, in game-theoretic settings. We have developed and analyzed a game-theoretic model with partially altruistic/spiteful players situated in a society or an economic/social network environment. In a long run, I am interested in working towards algorithmic game theory and mechanism design “in the real world”, currently including issues from a societal perspective such as fairness and reciprocity other than altruism and spite, as well as learning in repeated games. Broadly, I would like to apply algorithmic/mathematical analysis techniques and computational/experimental methods to tackle economic and social problems in general. In this proposal, I first summarize our results with extended questions that can be readily investigated in the first half year (Part I), and then propose more questions or directions that can be preliminarily explored in the first half year and fully investigated in the second half year (Part II).