In this paper, we present a brief overview of enterprise-wide optimization and challenges in multiscale temporal modeling and integration of different models for the levels of planning, scheduling and control. Next, we review Generalized Disjunctive Programming (GDP), as a new modeling paradigm for scheduling problems that are illustrated with the STN and RTN models. We then address scheduling problems that expand the scope of the area: simultaneous scheduling and heat integration, pipeline scheduling, crude oil and refined products blending, and demand side management. We illustrate the advantage of the GDP modeling framework, describe effective strategies for global optimization, and describe multistage affinely adjustable robust optimization for uncertain interruptible load. We address integration of planning and scheduling, for which several approaches are reviewed, including use of traveling salesman constraints for multiperiod refinery planning, and multisite planning and scheduling of multiproduct batch plants. We report computational results to highlight the challenges.
Bibliographical noteFunding Information:
Financial support from Fundação para a Ciência e Tecnologia through projects IF/00781/2013, UID/MAT/04561/2013, and the Center of Advanced Process Decision-making at Carnegie Mellon. Thanks also to Danielle Zyngier, Optimization Specialist from Hatch, for providing the fruit juice processing example used to illustrate UOPSS.
© 2018 Elsevier Ltd
- Demand side management
- Generalized disjunctive programming
- Mixed-integer programming