A New Design of Genetic Algorithm for Solving TSP

Abstract

In this paper, we develop an algorithm that is able to quickly obtain an optimal solution to TSP from a huge search space. This algorithm is based upon the use of Genetic Algorithm techniques. The algorithm employs a roulette wheel based selection mechanism, the use of a survival-of-the-fittest strategy, a heuristic crossover operator, and an inversion operator. To illustrate it more clearly, a program based on this algorithm has been implemented, which presents the changing process of the route iteration in a more intuitive way. Finally, we apply it into a TSP problem with fifty cities. By comparing with other published techniques, we can easily know that the proposed algorithm can efficiently complete the search process and derive a better solution.