Subarray-level Adaptive Beamforming Based on Multi-agent Genetic Algorithm

In order to solve the problem of subarray layout optimization in subarray-level adaptive beamforming, a subarray structure optimization technique based on multi-agent genetic algorithm (MAGA) is proposed. Based on the traditional genetic algorithm framework, this technology introduces Multi-agent cooperative search mechanism and differential evolution strategy enhance global search ability and convergence performance. The linear constrained minimum variance (LCMV) criterion is used to calculate the weighting coefficients corresponding to different subarray partition structures. The maximum sidelobe level and the null characteristics are combined to construct a weighted cost function to achieve the purpose of adaptive optimization of the subarray structure. The simulation results show that under the condition of non-uniform sub-array and easy engineering implementation, the adopted method can effectively reduce the sidelobe level of the pattern, and the maximum can be reduced to about -15dB. The effectiveness and stability of the method are verified by comparing the simulation results of different division methods and array sizes.