Herz A V
Physics of Computation Laboratory, California Institute of Technology, Pasadena 91125.
J Theor Biol. 1994 Jul 7;169(1):65-87. doi: 10.1006/jtbi.1994.1130.
A class of spatially extended evolutionary games with simple local rules is introduced. The emergent properties are studied through two complementary approaches. One is based on a heuristic local analysis, the other on exact global techniques. The local analysis provides criteria to group the games into classes with distinct behavior. The results facilitate numerical simulations and reveal that even simple games allow for complex spatio-temporal phenomena. The global analysis demonstrates that certain games perform an uphill march in a fitness landscape determined by the payoff parameters and the topology of the underlying lattice structure. For generic game parameters, the landscape is rugged owing to competing interactions and generates dynamical phenomena well known from frustrated systems: trapping in local maxima for noiseless dynamics and very long relaxation times for stochastic dynamics. Although the model is a mere caricature of evolutionary processes, some of its emergent properties are reminiscent of those observed in nature. It is argued that similar dynamical phenomena will be present in more elaborate approaches.
引入了一类具有简单局部规则的空间扩展进化博弈。通过两种互补的方法研究其涌现特性。一种基于启发式局部分析，另一种基于精确的全局技术。局部分析提供了将博弈分组为具有不同行为类别的标准。这些结果有助于数值模拟，并揭示即使是简单的博弈也会出现复杂的时空现象。全局分析表明，某些博弈在由收益参数和底层晶格结构的拓扑决定的适应度景观中进行上坡行军。对于一般的博弈参数，由于竞争相互作用，景观崎岖不平，并产生了受挫系统中众所周知的动力学现象：无噪声动力学中陷入局部最大值，随机动力学中弛豫时间非常长。尽管该模型只是进化过程的一个简单缩影，但其一些涌现特性让人联想到在自然界中观察到的特性。有人认为，在更精细的方法中也会出现类似的动力学现象。
