The Improvement of Ant Colony Algorithm Based on the Inver-over Operator

Abstract

As a typical paradigm of swarm intelligence, ant colony algorithm is another heuristic search algorithm applied in combinational optimized problem after simulated annealing algorithm, genetic algorithm, taboo search algorithm, ANN and so on. But ant colony algorithm has many shortages such as low rate while searching the solution and easy getting into a local optimization and so on. For these backwards, an improved ant colony algorithm based on the inver-over operator has been presented in this paper. In each circulation of ant colony algorithm, the optimum solution and the hypo-optimization solution have been operated 10 times by the inver-over operator. The algorithm can converge as quickly as possible and avoid getting into a local optimization. So it can increase the number of local solutions and enlarge the range of the optimum solution. The Improved algorithm has been simulated on travelling salesman problem (TSP) and compared with basic ant colony algorithm and another improved ACA based on cross operator. After the 100 times experiments, the comparison on average optimization and the proportion of optimization has shown that the algorithm proposed in this paper is better than the other two algorithms.