Flexibility Analysis in the Case of Incomplete Information about Uncertain Parameters

In recent years the flexibility analysis of chemical processes has attracted a significant amount of attention among researchers in the chemical engineering community. Flexibility analysis permits to identify/create chemical processes, which can satisfy all design specifications in spite of process...

Full description

Saved in:
Bibliographic Details
Published inAnnals of operations research Vol. 132; no. 1-4; pp. 257 - 275
Main Authors Ostrovsky, G.M., Achenie, L.E.K., Datskov, I.V., Volin, Y.M.
Format Journal Article
LanguageEnglish
Published New York Springer Nature B.V 01.11.2004
Subjects
Accuracy
Algorithms
Chemical engineering
Chemicals
Design
Design specifications
Feasibility
Flexibility
Guarantees
Operations research
Optimization
Parameter estimation
Process engineering
Studies
Variables
Online AccessGet full text

Cover

More Information
Summary:In recent years the flexibility analysis of chemical processes has attracted a significant amount of attention among researchers in the chemical engineering community. Flexibility analysis permits to identify/create chemical processes, which can satisfy all design specifications in spite of process and parametric uncertainty (from several sources) at the operation stage. All formulations of the flexibility problem are based on the supposition that during the operation stage there is enough experimental data from which exact values of the uncertain parameters can be obtained. However, in practice this assumption is often not met. Here in this paper, we consider the case when the uncertain parameters can be divided into two sets, namely a set that can be estimated with sufficient accuracy (at the operation stage) and a set that cannot be. Based on this view, we have developed extensions of the feasibility test and two-stage optimization problem to handle the two sets of uncertainty. We have developed the relevant split and bound algorithm for solving the new two-step optimization problem. [PUBLICATION ABSTRACT]
ISSN:0254-5330
1572-9338
DOI:10.1023/B:ANOR.0000045286.85889.db