Abstract: The Artificial Bee Colony algorithm ( ABC: Artificial Bee Colony) suffers from the problems of difficult convergence and difficulty in maintaining the diversity of candidate solutions. In order to solve the Multi- Objective Optimization Problem (MOP: Multi-Objective Problem) the solution strategy of each part is improved. Based on the ABC algorithm framework, a multi-objective ABC algorithm based on an adaptive solution strategy is designed to compare the performance of the improved multi-objective ABC with other typical swarm intelligence algorithms in the practical application of engineering design problem of electromechanical actuator design. The experimental verification shows that the proposed MOABC / DD(Multi-Objective Artificial Bee Colony Based on Dominance and Decomposition) algorithm has better problem solving accuracy compared with typical algorithms in solving the benchmark test case of electromechanical actuator design problem. The experimental results of MOABC / DD are more stable, thus proving that MOABC / DD has high solution stability and robustness.

Key words: multi-objective problem (MOP), artificial bee colony (ABC), swarm intelligence algorithms