Preprocessing algorithm and tightening constraints for multiperiod blend scheduling: cost minimization

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

While a range of models have been proposed for the multiperiod blend scheduling problem (MBSP), solving even medium-size MBSP instances remains challenging due to the presence of bilinear terms and binary variables. To address this challenge, we develop solution methods for MBSP focusing on the cost minimization objective. We develop a novel preprocessing algorithm to calculate lower bounds on stream flows. We define product dedicated flow variables to address product specific features involved in MBSP. Bounds on stream flows and new product dedicated flow variables are then used to generate tightening constraints which significantly improve the solution time of the mixed integer nonlinear programming models as well as models based on linear approximations.

Original languageEnglish (US)
Pages (from-to)603-625
Number of pages23
JournalJournal of Global Optimization
Volume77
Issue number3
DOIs
StatePublished - Jul 1 2020
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Management Science and Operations Research
  • Control and Optimization
  • Applied Mathematics

Keywords

  • Mixed integer nonlinear programming
  • Multiperiod blending
  • Preprocessing

Fingerprint

Dive into the research topics of 'Preprocessing algorithm and tightening constraints for multiperiod blend scheduling: cost minimization'. Together they form a unique fingerprint.

Cite this