北京大学学报(自然科学版)

合作进化模型综述

黄璜   

  1. 北京大学政府管理学院, 北京 100871;
  • 收稿日期:2009-12-21 出版日期:2011-01-20 发布日期:2011-01-20

Review on Model of Cooperation Evolution

HUANG Huang   

  1. School of Government, Peking University, Beijing 100871;
  • Received:2009-12-21 Online:2011-01-20 Published:2011-01-20

摘要: 作者综述了合作进化模型的最新研究成果。合作进化模型以“囚徒困境”博弈模型为基础, 为合作中的个体提供了一种与个体本身无关的策略优化机制, 以此探讨合作策略能够稳定进化的条件。基于直接或间接互惠理论的合作进化模型验证了在小规模社会中建立声誉机制有助于实现稳定的互惠合作; 互惠可以看做是对合作行为的奖赏, 开放的社会网络有助于实现较大范围内的合作; 对非合作行为实施惩罚有助于实现多人合作, 在一定条件下, 由非合作者实施惩罚时的合作进化相较于由合作者实施时更稳定; 群组竞争有利于合作文化或制度的传播, 但无法解释群组融合后的合作问题。

关键词: 合作进化, 互惠, 惩罚

Abstract: This article reviews current research on cooperation evolutionary model. Based on Prisoner’s Dilemma Game model, cooperation evolutionary model, offering individuals mechanisms of strategy optimizing unrelated with themselves, focuses on how cooperation can emerge from society. Evolutionary model based on theory of direct or indirect reciprocity proves reputation facilitate reciprocity cooperation in small scale society; reciprocity can be regarded as rewards to those cooperators that will dominate in society with open social network; punishment to those non-cooperators is good to n-person cooperation and a more stable equilibrium will emerge when non-cooperators are punishers under some conditions, comparing to when cooperators are; group competition facilitates the spread of cooperation culture or institution, but cannot explain cooperation when these groups merge.

Key words: evolution of cooperation, reciprocity, punishment

中图分类号: