A tabu search based metaheuristic for fast global optimizations of inverse problems

摘要

To develop a fast global optimizer, the whole iterative procedure of a tabu search algorithm is divided into two deliberately designed phases: exploration- and exploitation-phases. Stochastic approximation method is proposed to minimize the computational burdens when computing the gradient information in designing the exploitation phase. Also, some specially oriented mechanisms for enhancing the balance between exploration and exploitation searches are introduced and integrated. Numerical results are reported to showcase the merits of the proposed metaheuristic.