An adaptive weighting strategy for microwave imaging problems

In the framework of microwave imaging applications, an innovative strategy aimed at addressing the multiobjective inverse scattering problem is proposed. Starting from the spatial-domain integral formulation, the original multiobjective problem is recast into a single-objective one by defining a suitable cost function as a linear combination of the data and state terms according to variable weighting parameters. By iteratively tuning these parameters, the optimization procedure is forced to solve an "almost" multiobjective problem avoiding the use of ad-hoc multiple-objective optimization methods and satisfying different objectives in a balanced way. Selected numerical results indicate that the use of such a strategy yields to accurate reconstructions, with noise-corrupted data as well, by improving the performances of the adopted optimization procedure.