Abstract: For solving the Ant Colony System (ACS) own inherent defects, this paper proposes a novel combinatorial Ant Colony Optimization algorithm with detection zone strategy. In the proposed algorithm, the pheromone and the search path are modified dynamically. By using the detection method, the artificial ants are detected automatically per m iterations during the detection zone. When the ant colony falls into the local optimum, the variable q0 will be adaptively modified by the algorithm. Meanwhile, for improving the search abilities of artificial ants, it changes the global rate of pheromone evaporation and the maximum and minimum of pheromone, respectively. The performance of the novel algorithm is conducted, and the comparison among the original Ant System (AS), Ant Colony System (ACS) and the proposed algorithm is shown. The experiment result demonstrates that the CAAC(Combinatorial Adaptive Ant Colony) has a better performance than ACS in term of the capability of search and ability of restrain stagnation.

Key words: ant colony system, detection zone strategy, optimization algorithm