A Combined Algorithm Using Both the MINLP Model and Approximated MILP Model for PVC Production Scheduling

In this paper, a scheduling model of PVC production by a calcium carbide method is designed based on a continuous-time modeling method, and an improved mixed-integer nonlinear programming (MINLP) model for scheduling of PVC production is proposed. The optimization goal is to minimize the total cost. Considering the practical requirements of both the solution rapidity and quality, a combined algorithm is further established using both the MINLP model and approximated mixed-integer linear program (MILP) model for PVC production scheduling. The optimization result of the linear model is substituted into the original MINLP model as the initial value of variables to accelerate the solution process. Then, the optimal solution of the improved model is executed. Afterward, the effectiveness of the proposed method is verified with two actual cases. The comparative results demonstrate that the proposed algorithm can significantly accelerate the computation and obtain more accurate optimal solution.