Abstract:

In order to improve the convergence rate and preserve the diversity of the solutions to multi-objective numerical optimization problems, a quantum-inspired evolutionary algorithm is presented based on the principles of quantum computation and multi-objective optimization. In this algorithm, first, a crowed comparison operator is used to sort and select individuals according to the characteristics of multi-objective optimization. Then, a non-uniform mutation operator is applied to the observation populations to preserve the convergency of the solutions and to improve the precision of local search. Finally, a diversity-preserving operator is employed to delete the observation populations for the purpose of preserving the solution diversity. Experimental results show that, as compared with the NSGA-Ⅱ algorithm, the proposed algorithm is of higher convergence rate and better population diversity

Key words: numerical optimization, evolutionary algorithm, real-value coding, non-uniform mutation, square pulse