Improved Genetic Algorithm to Solve Asymmetric Traveling Salesman Problem

Abstract

The asymmetric traveling salesman problem (ATSP) is a combinatorial problem of great importance where the cost matrix is not symmetric, which complicates its resolution. The genetic algorithms (GAs) are a meta-heuristics methods used to solve transportation problems that have proved their effectiveness to obtain good results. However, improvements can be made by adapting the crossover operator as a primordial operator in GAs. In this work, we propose an adapted XIM crossover operator for the ATSP in order to improve the optimal solution obtained by GAs. Numerical simulations are performed and discussed for different series of standard instances showing the improvement of the optimal solution by the proposed genetic operator.