A new meta-heuristic approach for combinatorial optimization and scheduling problems

This study presents a new metaheuristic approach that reasonably combines different features of several well-know heuristics. The core component of the proposed algorithm is a simulated annealing that utilizes three types of memories, two short-term memories and one long-term memory. The purpose of the two short-term memories is to guide the search toward good solutions. While the aim of the long term memory is to provide means for the search to escape local optima through increasing the diversification phase in a logical manner. The long-term memory is considered as a population list. In specific circumstances, members of the population might be employed to generate a new population from which a new initial solution for the simulated annealing component is generated. Job shop scheduling problem has been used to test the performance of the proposed method.