Abstract
The optimization function of the method of search-space smoothing (SSS) in combination with the Metropolis algorithm (MA) is studied on the random Euclidean traveling salesman problem. The search dynamics of this combined approach (MASSS) are analyzed in both the smoothed and unsmoothed terrains and compared to those of the original SSS method. The results show that the MA can be successfully utilized as a local search algorithm in the SSS approach and the optimization characteristics of these two constituent methods are improved in a mutually beneficial manner in the MASSS run: The relaxation dynamics generated by employing the MA effectively work even in a smoothed terrain and more advantage is taken of the guiding function proposed in the idea of SSS. Also, this mechanism operates in an adaptive manner in the de-smoothing process and the MASSS method maintains its performance over a wider temperature range than does the MA.
