A Knowledge-Based Genetic Algorithm to the Global Numerical Optimization

Global optimization algorithms have received much attention recently. This paper presented a Knowledge-based Genetic Algorithm (KGA) for the global numerical optimization. In KGA, some innovative operators were proposed by integrating the empirical knowledge with the existing operation. In particular, we proposed two novel operators: knowledge-based mutation operator based on round or immunity operation, and knowledge-based local search operator based on sensitivity analysis and steepest descent method. The experimental results suggest that KGA outperforms to some published algorithms.