A novel modified gravitational search algorithm for the real world optimization problem

摘要

The directing orbits of chaotic systems is a common multimodal optimization problem in the engineering field. However, when this multimodal optimization problem is solved by evolutionary algorithm, it is difficult for the method to obtain the high-quality solution for easily falling into a local optimal solution. To address this concerning issue, a novel global gravitational search algorithm with multi-population mechanism (named GGSA) is proposed. GGSA makes use of the clustering method to divide the whole population into several subpopulations for maintaining the population diversity. Then, the information contained in global best agent is used to update the current agent for improving the convergence speed. By this way, the proposed algorithm can achieve a right tradeoff between the exploration and the exploitation. Finally, the directing orbits of discrete chaotic systems are used to test the performance of the proposed algorithm. The experimental results show GGSA has better performance than other compared methods.