Ant Colony Optimization for Intelligent Scheduling

摘要

A novel Ant Colony Optimization algorithm was proposed for the scheduling problem in multiproduct chemical batch process, including a critical block based neighborhood structure in local search procedure to reduce the searching space of the problem, an ants-seed strategy, a stagnation step out mechanism and a pheromone trail limit mechanism in pheromone updating procedure to avoid stagnation. In the proposed algorithm, the pheromone acts as an indirect communication media among the ant colony, guided by the pheromone, all the ants converge to good tours in the sence of probability. Comparisons with other well-performed algorithms on Taillard's benchmark problems show that our algorithm is more efficient and has stronger adaptability and robustness.