报告题目: Scheduling of Crude Oil Operations in a Refinery: Deterministic Formulation and Robust Optimization under Uncertainty Coupled with Global Optimization
报告人（单位）：李杰（中国科学院，过程工程研究所）
点评人（单位）：舒嘉（管理科学与工程系）；刘新旺（管理科学与工程系）
时间地点：2012年11月13日（周二）下午2点；九龙湖经管楼B201
Abstract:
Scheduling is a decision making process to determine when, where, how to produce a set of products given requirements in a specific time horizon, a set of limited resources, and processing recipes. The research area of scheduling of crude oil operations in a refinery has received great attention from both the academia and the industry in the past two decades. This is motivated by the increasing pressure to improve efficiency, reduce operating costs and increase profit margin arising from ever-changing crude prices, deteriorating crude qualities and growing environmental concerns.
Scheduling is a decision making process to determine when, where, how to produce a set of products given requirements in a specific time horizon, a set of limited resources, and processing recipes. The research area of scheduling of crude oil operations in a refinery has received great attention from both the academia and the industry in the past two decades. This is motivated by the increasing pressure to improve efficiency, reduce operating costs and increase profit margin arising from ever-changing crude prices, deteriorating crude qualities and growing environmental concerns.
In this talk, we will present an overview of the existing modeling approaches for scheduling problems and their important characteristics. Then, we will present our novel and efficient non-convex mixed integer nonlinear programming (MINLP) mathematical model for scheduling of crude oil operations in a refinery. This formulation incorporates many realistic operational features. To solve this non-convex MINLP formulation, we will present our spatial branch-and-bound algorithm with piecewise-linear relaxation strategy for global optimization. In addition, we will present our robust optimization framework to address some uncertainties such as demand fluctuations, and ship arrival delay during the real operations. Large-scale industrial case studies will illustrate the potential benefits of our proposed formulation, global optimization algorithm and robust optimization framework.
Bio: 李杰研究员是中国科学院“百人计划”入选者。曾在美国普林斯顿大学等高校从事研究工作，先后主持美国国家科学基金会 (National Science Foundation) 等重大科研项目。 主要研究方向属于过程系统工程学科，通过建立数学模型和开发数学全局优化算法解决化工物流管理中的实际问题。