An improved differential evolution for batching problem in steelmaking production

This paper studies a common production planning problem encountered in steelmaking production. This problem is to group different customer orders into a set of batches such that each batch satisfies the capacity and other technological constraints required by steelmaking furnace. We formulate the problem as a mixed integer programming problem by considering all practical technological requirements. To solve the problem effectively, we propose an improved differential evolution (DE) algorithm in which each individual of the population is represented by a real-coded matrix. To reduce the computation complexity, a parallel evolution mechanism is proposed to generate multiple populations for DE, in which each population evolves separately. Computational results on the practical production data show that the proposed algorithm can obtain better solutions as compared with other DE algorithms and the manual method. In addition, the proposed algorithm is also competitive in comparison with the optimization solver CPLEX.