Continuous ant colony optimization system based on normal distribution model of pheromone

A new method for the optimization of continuous problems is proposed based on the forage behavior of real ants, which extends ant colony system. The distribution of pheromone on the continuous space is simulated with normal distribution. A random generator is used with this distribution as the state transition rule to choose the next point to move to, and pheromone is updated by adjusting parameters of the distribution according to the transitions of ant colony. Ants aggregate gradually around the optimal food source under the direction of pheromone. Furthermore, the eugenic strategy and mutation strategy are incorporated to enhance the exploitation and exploration of ant colony, and hybrid continuous ant colony system (HCACS) is constructed. The experiments on optimization of test functions show that HCACS well fits for the optimization of continuous problems, and it gains comparative advantages when applied to high-dimension complex problems or problems with larger search space. HCACS involves fewer control parameters, which can be set easily.