“…Various numerical algorithms have been developed for solving inverse medium scattering problems, such as iterative methods like contrast source-type inversion methods [50,51], iteratively regularized Gauss-Newton method [34], least-squares methods [28], subspace optimization methods [11] and recursive linearization methods with multi-frequencies [5,6,14] and with single frequency [4], which can obtain satisfactory reconstructions while they are generally very time-consuming. Non-iterative methods such as reverse time migrations [10], singular sources methods [43], factorization methods [32,44] and sampling-type methods [8,15,23,26,42] are fast while their reconstructions may not be accurate. We refer to [9,12,20] for more extensive reviews on both theoretical analyses and numerical algorithms on inverse scattering problems.…”