Improved hybrid genetic algorithm and its application in auto-coloring problem

An improved hybrid generic algorithm is proposed to solve the problem of the computation complexity and the sensitivity of initial selected population in auto-coloring. The algorithm combines the local search ability of greedy algorithm and the global search ability of genetic algorithm, as well as the specialized knowledge of graph coloring, using tabu rules to avoid repeated coloring and local circulation caused by the crossover process of generic algorithm. The infeasible solutions are forbad, and the next generation population are optimized so the implement efficiency of the algorithm can be increased. The simulation results show that the hybrid genetic algorithm can solve the problem of coloring administrative region map efficiently, improve both the implement efficiency and the convergence speed.