Abstract: The artificial bee colony algorithm has good robust，high convergence speed and outstanding performance in global optimization, but its ability of local search is not good enough. In order to improve its ability of local search, an improved chaotic artificial bee colony algorithm was proposed. In the new algorithm, local search is executed nearby the mean of all individuals, selecting the better individual between the solution searched by chaos function and previous population. Experimental simulation shows that the improved algorithm not only accelerates the convergence rate, but also improves its accuracy, whose performance is better than that of the existing artificial bee colony algorithm.

Key words: artificial bee colony; chaos functions; local search