Improved simulated annealing algorithm for GTSP

By analyzing the principles and shortcomings of traditional simulated annealing algorithm, an improved simulated annealing algorithm for solving GTSP (Generalized Traveling Salesman Problem) is proposed. In this paper GTSP is changed into multi-segment map problem which is solved with dynamic programming algorithm by individual coding. Individual fitness function based on multi-segment map optimal sub-structure is designed to speed up algorithm calculation. Experimental test results show that the new algorithm has faster convergence and better solution than traditional simulated annealing algorithm.