The Value of Stochastic Modeling in Two-Stage Stochastic Programs with Cost Uncertainty

Erick Delage (), Sharon Arroyo () and Yinyu Ye ()
Additional contact information
Erick Delage: Department of Decision Sciences, HEC Montréal, Montréal H3T 2A7 Canada
Sharon Arroyo: Boeing Research and Technology, The Boeing Company, Seattle, Washington 98124
Yinyu Ye: Department of Management Science and Engineering, Stanford University, Stanford, California 94305

Operations Research, 2014, vol. 62, issue 6, 1377-1393

Abstract: Although stochastic programming is probably the most effective framework for handling decision problems that involve uncertain variables, it is always a costly task to formulate the stochastic model that accurately embodies our knowledge of these variables. In practice, this might require one to collect a large amount of observations, to consult with experts of the specialized field of practice, or to make simplifying assumptions about the underlying system. When none of these options seem feasible, a common heuristic has been to simply seek the solution of a version of the problem where each uncertain variable takes on its expected value (otherwise known as the solution of the mean value problem). In this paper, we show that when (1) the stochastic program takes the form of a two-stage mixed-integer stochastic linear programs, and (2) the uncertainty is limited to the objective function, the solution of the mean value problem is in fact robust with respect to the selection of a stochastic model. We also propose tractable methods that will bound the actual value of stochastic modeling: i.e., how much improvement can be achieved by investing more efforts in the resolution of the stochastic model. Our framework is applied to an airline fleet composition problem. In the three cases that are considered, our results indicate that resolving the stochastic model can not lead to more than a 7% improvement of expected profits, thus providing arguments against the need to develop these more sophisticated models.

Keywords: robust optimization; value of stochastic solution; mean value problem; regret minimization; fleet mix optimization; programming; stochastic; inventory; uncertainty (search for similar items in EconPapers)
Date: 2014
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (6)

Downloads: (external link)
http://dx.doi.org/10.1287/opre.2014.1318 (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:inm:oropre:v:62:y:2014:i:6:p:1377-1393

Access Statistics for this article

More articles in Operations Research from INFORMS Contact information at EDIRC.
Bibliographic data for series maintained by Chris Asher ().