Cooperative Learning Sensitive Agent System for Combinatorial Optimization

Systems composed of several interacting autonomous agents have a huge potential to efficiently address complex real-world problems. A new Learning Sensitive Agent System (LSAS) is proposed to address combinatorial optimization problems. Agents communicate by directly exchanging information and knowledge about the environment. Furthermore, agents of the proposed model are endowed with stigmergic behavior and are able to indirectly communicate by producing and being influenced by pheromone trails. Each stigmergic agent has a certain level of sensitivity to the pheromone allowing various types of reactions to a changing environment. For better search diversification and intensification, agents can learn to modify their sensitivity level according to environment characteristics and previous experience. The proposed LSAS model is tested for solving various instances of the Asymmetric Traveling Salesman Problem. Numerical experiments indicate the robustn ess and potential of the new metaheuristic.