Abstract:

The hybrid bee colony (HBC) algorithm is proposed to overcome the problems that the artificial bee colony (ABC) algorithm is easier to get local optimal solution and has poor local search capability and low accuracy. The combination of ABC algorithm’s local convergence and simulated annealing (SA) algorithm’s global convergence provides a new mechanism for the ABC algorithm. The algorithm integrates the information of concept of the global best solution’s the replacement value and individual optimum into the search equation of ABC algorithm. The selection method of substitution value is implemented through the Metropolis criterion of SA algorithm, namely, by regulating the temperature in probability to determine the replacement value of the global best solution. Then, according to the nonlinear to reduce the number of iterations to reduce the scouts search, so the search of the scout bees is also improved. This improvement can effectively improve the global search ability of the algorithm, and accelerate later convergence of the algorithm. The simulation results in eight complex numerical functions show that the proposed algorithm has better performances than traditional algorithms in search ability and accuracy.