Mixed Integer Linear Programming Based Approaches for Medium-Term Planning and Scheduling in Multiproduct Multistage Continuous Plants
Mixed Integer Linear Programming Based Approaches for Medium-Term Planning and Scheduling in Multiproduct Multistage Continuous Plants
复制标题
DOI：
10.1021/acs.iecr.6b04958
复制
发表时间：
2017-05
影响因子：
4.2
通讯作者：
A. Aguirre;Songsong Liu;L. Papageorgiou
中科院分区：
文献类型：
--
作者：
A. Aguirre;Songsong Liu;L. Papageorgiou
针对多产品多级连续生产线的计划与调度问题，提出了三种基于混合整数线性规划（MILP）的新模型。这些模型结合联合收割机的TSP（旅行商问题）制定的一般优先和单位特定的一般优先概念的主要思想，提供混合离散/连续时间表示的系统。此外，针对该问题的中等规模实例，推导出了一种包含滚动时域和迭代改进算法的高效求解方法。对不同模型参数的结果分析表明了新公式的优点和本文提出的求解方法的有效性。
This paper addresses the planning and scheduling problem of a multiproduct multistage continuous plant by three novel MILP-based (mixed integer linear programming) models. These models combine a TSP (traveling salesman problem) formulation with the main ideas of general precedence and unit-specific general precedence concepts to provide hybrid discrete/continuous time representations of the system. Also, an efficient solution approach involving rolling horizon and iterative-improvement algorithm is derived for solving medium-size instances of the problem. Results analyses for different model’s parameters demonstrate the benefits of the new formulations and the effectiveness of the solution approach presented in this work.