Adaptive Maxwell's equations derived optimization and its application in antenna array synthesis

In this paper, a self-adaptive method for the Maxwell's Equations Derived Optimization (MEDO) is proposed. It is implemented by applying the Sequential Model-Based Optimization (SMBO) algorithm to the iterations of the MEDO, and achieves the automatic adjustment of the parameters. The proposed method is named as adaptive Maxwell's equations derived optimization (AMEDO). In order to evaluate the performance of AMEDO, eight benchmarks are used and the results are compared with the original MEDO method. The results show that AMEDO can greatly reduce the workload of manual adjustment of parameters, and at the same time can keep the accuracy and stability. Moreover, the convergence of the optimization can be accelerated due to the dynamical adjustment of the parameters. In the end, the proposed AMEDO is applied to the side lobe level suppression and array failure correction of a linear antenna array, and shows great potential in antenna array synthesis.