Ant Colony Optimization (ACO) system is an intelligent multi-agent system of the interacting artificial ants to solve the combinatorial optimization problems. Applying ACO approach in the typical NP-hard problem like job shop scheduling (JSS) problem is still an impressive and attractive challenge with the community. This paper proposes two improvements of ACO algorithm based on the convergence property of pheromone trails. Our improvements are better in both terms of accuracy and running time than the state-of-the-art Max-Min ant system by the simulation with the standard data sets. © 2008 Springer Berlin Heidelberg.

Do Duc D., Dinh H.Q., Hoang Xuan H.

414.pdf    Gửi cho bạn bè

Từ khóa : Agents; Artificial intelligence; Combinatorial mathematics; Combinatorial optimization; Convergence of numerical methods; Electric load forecasting; Intelligent agents; Optimization; Scheduling; ACO algorithms; ACO convergence; Agent systems; Ant colony optimization (ACO); Ant colony optimization algorithm; Ant system (AS); Convergence properties; Job shop scheduling; Job shop scheduling problem; NP hard problems; Optimiz ation problems; Pheromone trails; Running time; Update rules; Multi agent systems