Evolutionary Algorithm Based on Neutral Evolution, Self-Organization and Natural Selection
Xiang Wang · Shuju caiji yu chuli · 2003
Biological evolution takes place in genetic, individual and group levels. Genetic evolution is random, even and non directional. The random actions of individual form the complicated and ordered actions of group by self organization and the group evolution are a natural selection process under the pressure of environments. A novel evolutionary algorithm, based on neutral evolution, self organization and natural selection (NSNEA) is proposed. Evolution of gene, individual and group are all considered in the algorithm. Meanwhile, it contains the interaction and the map among the above three ones. Individual capability evaluation function f(x i) is presented, then the relation between f(x i) and group fitness function fit(X) is discussed. The performance of the algorithm is analyzed by simulation. Simulational result confirms that NSNEA is better than traditional evolutionary one in the aspects of global optimization, convergence speed and the robustness.