標題: 社會機制下多人囚犯困局理論中合作的演化
The Evolution of Cooperation: On the N-player Prisoner's Dilemma with Social Norms
作者: 陳宥任
Ru-Ren Chen
孫春在
Chuen-Tsai Sun
資訊科學與工程研究所
關鍵字: 囚犯困局;多人囚犯困局;社會機制;名聲;區域性;基因演算法;Prisoner's Dilemma;N-player Prisoner's Dilemma;Social Norm;Reputation;Spatial;Genetic Algorithm
公開日期: 2002
摘要: 囚犯困局的出發點,便是觀察個體的利己主義對於相互之間的合作關係所造成的影響。囚犯困局所衍生的理論已廣為應用在社會學,經濟學,及生物學等領域中。但是隨著環境的演化,社會已愈趨向複雜化。現實世界中的社會及經濟問題,並不僅僅限於個體與個體之間的互動,因此產生了多人囚犯困局的模型來模擬更接近現實社會中合作行為的表現。 在社會學中,能夠有效節制私人濫用便是促進整體利益的一大因素。因此在多人囚犯困局中,促進整體的合作行為是個很重要的研究方向。而在多人囚犯困局的研究中顯示,現實社會中的規模愈大,要促進整體合作行為就愈加困難。因此以社會機制來影響多人囚犯困局,以達到整體行為趨向合作,並觀察個體的策略在此環境下是否能達到收斂便是此研究方向的手段與目的。 在理論上條件式的偏見對手選擇會促進團體中合作的產生。而在陌生環境中,「以標籤為基礎的偏見對手選擇」是有效促進合作的手段。在熟悉的族群中,同伴之間的影響力則是現實中不容忽視的。本篇論文的主要研究方向便是試圖以「名聲」及「團體關係」,模擬現實中這些社會機制對群體的影響,以促進多人囚犯困局中合作的浮現。
Observing the effect of individualism on cooperation is the motivation behind the famous model of Prisoner’s Dilemma. The Prisoner’s Dilemma model can be applied in many fields, such as sociology, economic and biology. However, along with the evolution of our environment, the social structure becomes more and more complicated. The social and economic problems are no longer resulted from only the interactions between two individuals. Therefore, the N-player Prisoner’s Dilemma (NPD) model, which is used to simulate the cooperation in more structured world, has been proposed and studied. Based on experience from social reality, it has been seriously considered to increase the total benefit of the whole population by restricting the individual overuse of shared resource. Thus, promoting cooperation becomes an important issue in NPD study. Nevertheless, the outcome of NPD experiment shows that the larger the population-size, the more difficult to promote cooperation. Therefore, to promote the cooperation in the NPD framework by adding social norms becomes the focus of this research, We try to observe if individual tactics will converge into collective cooperation in an N-player environment. Theoretically, it encourages stronger cooperation in the group when biased opponent selection is used. In an unknown environment, the best way to promote cooperation is by tag-based opponent selection. While in a familiar circle, the influence from peers is obvious. The main research goal of this thesis is to simulate how “reputation” and “group-relationship” affect the promotion of cooperation in a structured world.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT910394036
http://hdl.handle.net/11536/70208
顯示於類別:畢業論文