A hybrid genetic algorithm for solving the economic lot scheduling problem (ELSP)

In this paper, a hybrid genetic algorithm(HGA) is proposed to solve the ELSP. The ELSP is formulated using basic period(BP) approach. In the proposed HGA, the design of basic setups of GA is largely adopted from previous contributions. Four heuristic strategies are introduced in the proposed HGA, aiming at reaching feasible space easier, testing feasibility and avoiding falling into local optimum. To evaluate the performance of the proposed HGA, we apply it to Bomberger´s classical problem under 88% and 66% utilizations. Our experiments indicate that the proposed HGA outperforms traditional GA in getting minimum total cost and faster convergence rate. More importantly, our proposed HGA is showed to be capable of converging to global optimal solutions.