Stauffer D, Jan N
Physics Department, Saint Francis Xavier University, Antigonish, N.S., Canada.
J Theor Biol. 1994 May 21;168(2):211-8. doi: 10.1006/jtbi.1994.1100.
Millions of sites are simulated in an NK fitness model of evolution and coevolution. We find a logarithmic size dependence of the number of hill-climbing iterations needed to reach a local fitness optimum (Nash equilibrium). We also check for chaotic behavior and determine the size of the damage clouds or avalanches. Random noise (simulated annealing) is shown to increase appreciably the fitness.
在进化与共同进化的NK适应度模型中模拟了数百万个位点。我们发现，达到局部适应度最优值（纳什均衡）所需的爬山迭代次数与规模呈对数关系。我们还检查了混沌行为，并确定了损伤云或雪崩的规模。结果表明，随机噪声（模拟退火）能显著提高适应度。
