Abstract: By analyzing the optimization mechanism of artificial bee colony algorithm, an improved artificial bee colony algorithm is proposed in terms of the initial population construction, subpopulations grouping, step updating and population elimination. The new algoBy analyzing the optimization scheme of the artificial bee colony algorithm, an improved version of such an algorithm is proposed in terms of the initial population construction, subpopulations grouping, step updating and population elimination. The new algorithm constructs the initial population by using the uniform design theory and a Z-type grouping method based on cross population is proposed. Specifically, an adaptive step based on logarithmic functions is designed to replace the original random step. At the same time, the population elimination mechanism based on niche technology is adopted to eliminate these individuals which have fallen into the local optimum in time. Experimental results show that the improved algorithm can avoid premature convergence, accelerate the searching rate and improve the accuracy of the solution.

Key words: artificial bee colony algorithm, uniform design, Z-type grouping, adaptive step, niche