Research of Combinatorial Optimization Problem Based on Genetic Ant Colony Algorithm

Ant colony algorithm of the traditional combinative optimization consumes a large amount of time in the process of solving the optimization, which has a tendency to partial optimization and slow convergence along with many redundant useless iterative codes and low operation efficiency. A generic optimized ant colony algorithm is thus proposed. This algorithm has the ability to fast global search of generic algorithm along with parallelism and positive feedback mechanism of ant algorithm. It determines the distribution of pheromone on the path by means of generic algorithm changing selection operators, crossover operators and mutation operators. Then ant algorithm is applied into feature selection. Supporting vector machine classifiers is used to evaluate the performance of the feedback sub-variorum. The pheromones are recombined through changing the pheromone iteration, parameter selection and increasing the local update of pheromones feature nodes. The simulation experiment shows that this algorithm can improve the accuracy effectively, speed up the convergence, improve global optimization, and promote the robustness and stability.