A Genetic Algorithm Design Based on Self-Organizing Dynamic Network

Abstract

In order to improve the population diversity and convergence performance of genetic algorithm, a self-organizing dynamic network model is introduced into the neighborhood structure of genetic algorithm. In order to evaluate the importance of network nodes more completely and effectively, a new definition of exponential network node fitness is given firstly, which considers the ranking of the objective function value of nodes in neighbor nodes and the number of neighbor nodes. Then, three kinds of topology updating rules, i.e. double production, single production and selective deletion, are proposed to make the network topology evolve dynamically with the evolution of genetic algorithms. Test results of these typical optimization functions show that the genetic algorithm designed in this paper is superior to standard genetic algorithms and small-world genetic algorithms in population diversity and convergence performance.