A novel geometric center design method for genetic algorithm optimization

This paper presents a novel geometric center embedded genetic algorithm (GCEGA) in solving optimization problems. Due to the inherent characteristics of genetic operators, traditional genetic algorithm (GA) has weaknesses in exploiting a peak. Hence it is time-consuming to attain a high precision solution. To deal with this problem, the geometric center design (GCD) method is proposed. It utilizes the geometric knowledge to approach the geometric center (GC) in search of potential optimum values. In every generation, some high-quality individuals are chosen to compute the GC, which is then evaluated and conditionally put back into the population. Experiments have been implemented on twelve functions for comparison between the traditional GA and the proposed algorithm. The results reveal that the proposed algorithm can remarkably enhance the performance of the traditional GA with faster speed and higher accuracy.