Genetic Algorithm-based Chaos Clustering Approach for Nonlinear Optimization

In this paper, a n-dimension convergence algorithm was employed to track the potential trend of evolution in traditional genetic algorithm (GA) by K-means clustering technique. And, chaotic algorithm was exploited to prevent the new approach from premature. By means of the proposed approach, not only the basic search capability was maintained but also the flexibility and efficiency of parametric modeling were improved. The main purpose of the paper is to demonstrate how the GA optimizer can be improved by incorporating a hybridization strategy. Experimental studies revealed that the hybrid chaotic approach with genetic algorithm (CGA) procedure could produce much more accurate estimates of the true optimum points than other optimization procedures. Furthermore, including K-means clustering into CGA, named KCGA, exhibited superior convergence performance than other algorithms. And, the proposed approach, KCGA, had 84 percent of probability to get optimized. On the whole, the new approach was demonstrated to be extremely effective and efficient at locating optimal solutions and verified by an empirical example from construction.