In this paper

we propose an improved evolutionary algorithm combining diversity maintaining mechanism and accelerating operators

which focuses on the contradiction between the maintenance of population diversity and search efficiency in solving multimodal function global optimization problem on a bounded area.The convergence analysis shows our algorithm can converge to global optimization under some circumstance.When dealing with low dimensional problems

our algorithm is prone to converge to global optimization and outperforms canonical genetic algorithm

while with high dimensional problems

the converging condition cannot be satisfied

our algorithm can converge to the so-called ε-satisfied result and outperforms canonical particle swarm optimization algorithm.