An improved multi-objective differential evolution algorithm and its application for array pattern synthesis

An efficient improved multi-objective differential evolution algorithm (IMODE) is presented in this paper. By utilizing an improved mutation strategy with the concept of random tournament selection, the proposed algorithm generates a well distributed Pareto optimal front. Then, the IMODE is applied in an array pattern synthesis problem in which a smaller beam width and lower side lobe level (SLL) with null in certain direction is necessary. The results show the IMODE has a better performance than NSGA-II and GDE3 in convergence to Pareto front and diversity of solutions, which indicate that the proposed algorithm is a competitive algorithm in multi-objective optimization problems.