A hybrid orthogonal genetic algorithm for global numerical optimization

In this paper, a hybrid orthogonal genetic algorithm (HOGA) is presented to solve global numerical optimization problems of continuous variables. Based on traditional genetic algorithms, the HOGA has been augmented with a robust selection operator and an intelligent crossover operator. These augmentations reduce statistical bias while improving convergence times and relative accuracy of the solutions. Examples show that HOGA can effectively solve a number of multimodal problems which are widely accepted as optimization benchmarks.