An Improved Pareto Genetic Algorithm for Multi-objective TSP

Multi-object traveling salesman problem (MOTSP) is a typical multi-object optimization problem. It requires to select a best route and make a balance between cost assignment and distance assignment of the route, the less cost of the whole travel and to satisfy the stipulate is the guide line. This paper gives the non-domination of genetic algorithm, and shows a simple model to put out the method that using multi-object genetic algorithm to solve the TSP. The algorithm use integer coding method, create an initial population that satisfies the basic qualification; calculate the two objective-value: distance and cost; then rank the chromosomes with Pareto function according to the objective-value; and use tournament selection to select the better chromosomes to form a series of parents, through multi-objective greedy crossover, and then use transposition mutation algorithm; we can get a new population that forms of new individuals based on genetic-searching function, and get the approximately best solution at last. The computing results of real examples of the MOTSP demonstrates that the approximate global optimal solution of the problem can be quickly obtained, and the solution with high accuracy.