[[alternative]]Tests on Genetic Algorithm for Traveling Salesman Problem

[[abstract]]The traveling salesman problem (TSP) is an NP-complete problem. These problems are considered to have no solution in polynomial time. However, there are ways of approximating the solutions to these problems. Genetic algorithm (GA) is a model of machine learning based on biological evolution of population, which uses crossover and mutation operators to solve optimization problems, evaluate individual by fitness function. They have been used successfully in different problems, including the TSP. In this paper, we design a restriction action to weed out some paths. It is a new idea. Restriction had better be a necessary condition of the optimal solution, but they also work even though they are not. Restriction itself is not an operator. This can be added to other operators to make them more efficient. We also compare with some studies of other researchers. We find that a suitable restriction can increase the efficacy of GAs. We can design other restrictions on GAs for TSP or other problems.