STUDY OF BIOTIC EVOLUTIONARY MECHANISMS BASED ON THE MULTI-AGENT PARRONDO'S GAMES

2012 
The paper tries to explain the biological evolutionary mechanism based on Parrondo's games. The multi-agent Parrondo's games proposed include game A and game B. Game A, a zero-sum game, reflects the interactive relationship of competition or cooperation between individuals. A two-dimensional lattice network is used to represents the spatial carrier of interactive relationship. Game B, a negative-sum game, reflects the environmental effect on survival evolution of individuals. Three patterns which are based on individual survival states, on historical experience and on neighbors' environment are adopted to represent the mechanisms. The simulation results show that: (1) A zero-sum game (game A) and a losing game (game B) combined could produce a winning result. Improvement of the average fitness of the population represents the essence that Parrondo's paradox is counterintuitive. The special structure of game B reflects the ratcheting effect that natural environment has on biological evolution where both competition and cooperation are successful evolution directions. (2) The ratcheting effect of the environment-based pattern is the strongest. In the course of biological evolution, neighbors' environment usually uses an efficient mechanism to guide the evolution. Just because of the neighbors' environment-based guidance, a coevolution model is formed where individuals compose the population which composes the community and the community composes the ecosystem which composes the biosphere. (3) Competition can raise the average fitness of the population. Therefore, competition may be an adaptive behavior on the population level. (4) Research to-date has shown that the necessary condition of producing cooperation is that the cost is less than profit. While the results in this paper show that a zero-sum game can also give rise to cooperation, which can bring about positive gain for the whole population.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    25
    References
    6
    Citations
    NaN
    KQI
    []