MMAS Based on Grey Prediction and Cloud Association Rules

Aiming at the shortcomings of slow convergence speed and easily trapping in local optimum in ant colony algorithm (ACO), an improved version of basic max-min ant system (MMAS) is proposed. It uses the grey information which is obtained form pheromone matrix each iteration to build grey model to predict and control the maximum and minimum trail limits real-timely. Meanwhile, it also makes some other parameters of algorithm controlled by using cloud association rules. Through both of the improved strategies, the algorithm can avoid effectively the slow convergence caused possibly by implementing the max-min trail limit strategy and the early stagnation of search, and appease the contradiction between the convergent speed and the searching scope dynamically. The simulation result for JSP shows the validity of it.