Application of the simulated annealing algorithm to the combinatorial optimization problem with permutation property: An investigation of generation mechanism

Focusing on the generation mechanism of random permutation solutions, this paper investigates the application of the Simulated Annealing (SA) algorithm to the combinatorial optimization problems with permutation property. Six types of perturbation scheme for generating random permutation solutions are introduced. They are proved to satisfy the asymptotical convergence requirements. The results of experimental evaluations on Traveling Salesman Problem (TSP), Flow-shop Scheduling Problem (FSP), and Quadratic Assignment Problem (QAP) reveal that the efficiencies of the perturbation schemes are different in searching a solution space. By adopting a proper perturbation scheme, the SA algorithm has shown to produce very efficient solutions to different combinatorial optimization problems with permutation property.