Research on Steelmaking Production Scheduling Method Incorporating Benders Decomposition

Abstract

The steelmaking-continuous casting production scheduling problem is a typical multi-stage, multi-constraint hybrid flow shop problem. For this problem, a two-stage planning method based on benders decomposition is proposed, and the problem is decomposed into two subproblems: batch planning scheduling based on pouring times and time preparation scheduling based on furnace times, and a mathematical linear programming model based on a specific unit event point is established with the objectives of minimizing the total processing time of the casting machine and minimizing the maximum completion time, respectively. For problem solving, a two-stage genetic algorithm based on bender decomposition is designed, combined with the parallel backward preparation method, to realize the scheduling of pouring schedule in the continuous casting stage and the processing schedule of furnace in each process, respectively. Finally, simulation experiments based on actual steel production data show that the proposed two-stage planning method based on bender decomposition can effectively solve this production scheduling problem.