Integrated Transfer Planning and Scheduling for the Material Blending Production System

Abstract

In this thesis, we study an integrated optimization problem for the production system where materials are transferred by paired containers from preprocessing plants to blending plants, and products are manufactured by blending materials at blending plants. The problem is formulated as a mixed-integer linear programming model considering the machine environment of the blending plants, due dates and required component concentrations of products, and ready times of materials. The purpose of the model is to determine container pairings, blending plant assignments, and schedules for blending operations to minimize the total accomplishment time of using containers, the total penalty for components outside the required component specification, and the workload of employees. In order to obtain good approximate solutions, a hierarchical decomposition strategy is proposed, and a heuristic approach and a Benders decomposition algorithm are developed for each decomposed problem. We show that applying the additional cut generation and neighbor solution search methods to a Benders decomposition algorithm can yield a computational time effect through numerical experiments. Furthermore, we demonstrate that our proposed methods can find good approximate solutions in a shorter time than solving a mixed-integer linear programming formulation by a commercial optimizer.