Linear Array Pattern Synthesis Using Multi-Objective Optimization Algorithm Based on Reference Vectors

In this paper, non-dominated sorting genetic algorithm based on reference vectors (NSGA-II/BRV) is proposed to minimize the Peak Sidelobe Level (PSL) and the Average Broad Null Level (ABNL) for the pattern synthesis of linear array. Here, NSGA-II/BRV mainly focuses on the convergence and diversity of the obtained solutions for multi-objective problems. In order to get a better distribution and convergence of the obtained solutions, the members close to the uniformly distributed reference vectors will be selected. Compared with MOPSO, numerical examples demonstrate that NSGA-II/BRV applied to array pattern synthesis can obtains higher performance and better distribution of the final solutions.