Reinforcement learning enhanced multi-neighborhood tabu search for the max-mean dispersion problem (Q2673245)

The paper presents an enhanced methodology to solve the max-mean dispersion problem, as a combination of multi-neighborhood tabu search and reinforcement learning. A Q-learning mechanism uses the response from the search to diversify the solutions and a learning automata is integrated for an adaptative probability choice for neighborhood selection. An analysis of the performance of the proposed technique is achieved by comparison with state-of-the-art methods on 80 test scenarios: the best results were also found by the new approach, while new lower bounds were discovered for 3 instances.