A Bionics Evolutionary Algorithm Based on Average Vector Deviation

Through an analysis of defects of the traditional evolutionary algorithm, a bionics evolutionary algorithm, based-on average vector deviation (BEAVD), is proposed in this paper. This new algorithm greatly improves the traditional evolutionary algorithm and has many good features. It simulates the queen-bee´s method of mating in nature. What´s more, aiming at enhancing the diversity of population and ultimately getting the global solution, BEAVD also introduces the strategy of Guo´s algorithm(GuoA), uniform recombination(UR) as well as the average vector deviation mutation factor(AVD). In numerical experiments, we have used this algorithm to solve some complex single-objective optimization problems with high complexity, which the traditional evolutionary algorithm and even the GA and ES have trouble to solve. Compared with the traditional evolutionary algorithm, BEAVD has better performance in realizing global optimization and promoting evolution efficiency, and can lead to higher precise solutions.