Adaptive Genetic Algorithm with Application for Solving Traveling Salesman Problems

In genetic algorithm (GA), 2-opt, a local search mechanism, is often used as as mutation operation for solving Traveling Salesman Problems (TSPs). The 2-opt mutation exchanges some edges of parents to generate new children. However, the operation of random edge inversion destroys some good edges in later convergence process. To solve this problem, an adaptive 2-opt operation is proposed. The probability of sub path inversion is reduced gradually with the process of convergence. The new method has been used to solve the Traveling Salesman Problems. Experimental results indicate that the adaptive GA is superior to the classical ones.