Two-stage robust optimization, state-space representable uncertainty and applications. Issue 4 (10th June 2014)

• Record Type: Journal Article
• Authors: Minoux, Michel
• Is Part Of: RAIRO. Volume 48:Issue 4(2014)
• Page Start: 455
• Page End: 475
• Publication Date: 2014-06-10
• Subjects: Robust optimization, -- Graph algorithms, -- Min-Max optimization; Operations research -- Periodicals; 658.4034
• DOI: 10.1051/ro/2014017
• Languages: English
• ISSNs: 0399-0559

Abstract

Abstract : The present paper addresses the class of two-stage robust optimization problems which can be formulated as mathematical programs with uncertainty on the right-hand side coefficients (RHS uncertainty). The wide variety of applications and the fact that many problems in the class have been shown to be NP-hard, motivates the search for efficiently solvable special cases. Accordingly, the first objective of the paper is to provide an overview of the most important applications and of various polynomial or pseudo-polynomial special cases identified so far. The second objective is to introduce a new subclass of polynomially solvable robust optimization problems with RHS uncertainty based on the concept of state-space representable uncertainty sets . A typical application to a multi period energy production problem under uncertain customer load requirements is described into details, and computational results including a comparison between optimal two-stage solutions and exact optimal multistage strategies are discussed.