Ver Post

Seminário do CEG-IST

22 setembro 2017, 11:12 - Fátima Jacinto

No âmbito do Ciclo de Seminários do CEG-IST, e do projeto DM4Manufacturing, temos o prazer de anunciar o seguinte seminário:

 

Data: 27 de Setembro de 2017 (4.ª feira), das 11:00 às 12:00

Local: Sala Q4.7 (Técnico, Campus Alameda, Torre Sul, Piso 4)

Tema: From Scheduling to Online Scheduling: Models, Solution Methods, Paradoxes, and Open Questions

 

Palestrante: Prof. Christos T. Maravelias - Vilas Distinguished Achievement Professor and Executive Officer of the Department of Chemical and Biological Engineering at the University of Wisconsin – Madison.

His research interests lie in the areas of (1) chemical production planning and scheduling; (2) chemical process synthesis and analysis; (3) assessment of renewable technologies for fuels and chemicals; and (4) computational methods for material discovery. 

 

Resumo: The talk consists of two parts: (1) an overview of recent progress in “traditional” chemical production scheduling, and (2) a discussion of the dynamic nature of scheduling and how to perform it repetitively.  

In the first part, we start with an overview of the different types of chemical production scheduling problems and the mixed-integer programming (MIP) models that have been proposed to address them. We then briefly discuss three solution methods for MIP scheduling models: (1) algorithms for the generation of unit-, task-, material-, and resource-specific nonuniform grids used to formulate multi-grid discrete-time models; (2) tightening inequalities based on a constraint propagation algorithms; and (3) reformulations and branching strategies. We present computational results showing that the proposed methods lead to substantial computational improvements. 

In the second part, we first and foremost show that open-loop and closed-loop scheduling are two different problems, even in the deterministic case. Second, when uncertainty is present, we show that acting on new information can in fact be more important than acting on observed uncertainty, and suggest that it is important to reschedule periodically, even when there are no “trigger” events, something that is in contrast with the current approaches. Thus, we propose the term “online scheduling” to describe the process of re-optimization to react to uncertainties/disturbances as well as new information arrival. Next, we discuss how suboptimal open-loop solutions can be corrected through revisions due to feedback and we study how rescheduling frequency and moving horizon length of the open-loop problem affect the quality of closed-loop schedules. We close with a set of open questions in the area of online scheduling.

 

Os seminários do CEG-IST são de acesso livre.