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.