In allusion to the problems that the traditional genetic algorithm has slow convergence velocity and is liable to be trapped in local convergence, an improved genetic algorithm with heuristic search method is proposed. In addition, the probabilities of mutation can be adaptively adjusted. Grid method and 0-1 matrix are used to establish work environment model, the heuristic search method with direction selection and the wall following strategy is used to produce the initial population, avoiding producing invalid code and decreasing the search area. As a key point, the effect of two kinds of adaptive mutation rates on algorithm is discussed in order to improve the performance of the algorithm. The mechanism of variable mutation rate and saving the best individual is used. In addition, a new modify operator is adopted to optimize individual. The simulation result proved that the optimal path is found successfully in a short time and the algorithm has better performance. http://www.sj-ce.org/paperInfo.aspx?ID=641
