The investigation of the regularization term in the continuous-parameter subspace based optimization method in reconstructing PEC objects

This paper investigates the influence of the regularization term on the performance of the continuous-parameter subspace based optimization method in reconstructing perfectly electric conductors (PEC), which provides a theoretical explanation to the choice of regularization term. The number of leading singular values of a current-to-field mapping operator is found to balance the accuracy and the stability of the solution. If the number of leading singular values is chosen as a large number, the noise is amplified in the inverse process. On the other hand, if this parameter is chosen to be a small number, the convergence of the optimization method will be slow. It is found that the regularization term in the PEC SOM behaves similarly as its counterpart in the dielectric case.