摘要The emergence of cooperation and the effectiveness of penalties among competing agents are studied via a model of evolutionary game incorporating adaptive behavior and penalties for illegal acts. For initially identical agents, a phase diagram is obtained via an analytic approach, with results in good agreement with numerical simulations. The results show that there exists a critical penalty for achieving a completely honest population and a sufficiently well-behaved initial population, requires no penalty. Self-organized segregation to extreme actions emerges in the dynamics fora system with uniformly distributed initial tendencies for cooperation. After training, the penalty can be relaxed without ruining the adapted cooperative behavior. Results of our model ill a Population taking oil the form of a 2D square lattice are also reported. (C) 2009 Elsevier B.V. All rights reserved.

著者Xu C, Hui PM, Yu YY, Gu GQ

期刊名稱Physica A: Statistical Mechanics and its Applications

出版年份2009

月份10

日期15

卷號388

期次20

出版社Elsevier

頁次4445 - 4452

國際標準期刊號0378-4371

電子國際標準期刊號1873-2119

語言英式英語

關鍵詞Penalty; Self-organized cooperation

Web of Science 學科類別Physics; Physics, Multidisciplinary; PHYSICS, MULTIDISCIPLINARY