We had the pleasure to host Dr. Pedro Castro (University of Lisbon, Portugal) at Aalto University for a week in November. During the visit, Dr. Castro gave an open seminar, entitled "Optimization Models for Scheduling in the Process Industries".

Abstract: Scheduling is a key element in enterprise-wide optimization, which concerns optimizing the supply, manufacturing and distribution activities of a company. Scheduling problems typically deal with time horizons of days or weeks and are typically tackled using mixed-integer linear (MILP) or nonlinear (MINLP) models. One of the challenges in scheduling is the wide variety of problems that can be found, which has hindered the development of a model/algorithm/commercial software that is both general and computationally efficient. Nevertheless, major developments have occurred over the last 30 years in the Process Systems Engineering community, which will be briefly reviewed. Examples include the unified frameworks for process representation (State- and Resource-Task Network), discrete- and continuous-time formulations, and the use of Generalized Disjunctive Programming to help formulating difficult constraints. We will then discuss 6 eal-life problems from the chemical and petroleum industry. They deal with time-dependent electricity costs, heat integration in batch plants, transportation of liquids by pipeline, refinery planning and crude oil blending.

The slides of the seminar can be found here.