Risk Analysis 101 --- Robust-Optimization: the elephant in the robust-satisficing room
Moshe Sniedovich (moshems.unimelb.edu.au)
Abstract: In 2001, info-gap decision theory re-invented the then 40-year old model of local robustness, known universally as radius of stability (circa 1960). Since then, this model of local robustness has been promoted by info-gap scholars as a reliable tool for the management of a severe uncertainty that is characterized by a vast (e.g. unbounded) uncertainty space, a poor point estimate of the uncertainty parameter and a likelihood- free quantification of uncertainty. Inexplicably, this absurd proposition has managed to pass muster in the review processes of academic books and journals. Small wonder then that info-gap’s robustness model was subsequently proposed, in a peer-reviewed article, as a framework for dealing with Taleb’s Black Swans and even . . . Unknown Unknowns?! More recently, the promotion of info-gap decision theory from the pages of peer- reviewed journals, has been conducted under the more general banner of the great merit of the robust-satisficing approach in decision-making, to the effect that advocates of this theory are now engaged in re-inventing the field of Robust Optimization. The trouble in all this is that the misguided rhetoric on robust-satisficing coming out of the info-gap literature, specifically the misguided rhetoric on the advantage of satis- ficing over optimizing that ends obscuring the obvious connection of robust-satisficing to Robust Optimization, had not been recognized for what it is, in the review process of peer-reviewed journals, such as Risk Analysis. In view of this state-of affairs, my objective in this discussion is to make it abundantly clear that info-gap’s robust-satisficing approach is in fact a step backwards to the early days of Robust Optimization, a step that completely ignores the tremendous progress over the past forty years or so, in this highly active area of research.
Keywords: Robust optimization, decision theory, radius of stability, maximin, local and global robustness, robust-satisficing, info-gap decision theory.
Category 1: Robust Optimization
Citation: Working Paper SM-12-1, July 2012, Department of Mathematics and Statistics, The University of Melbourne, Melbourne VIC 3010, Australia
Entry Submitted: 07/04/2012
Modify/Update this entry
|Visitors||Authors||More about us||Links|
Search, Browse the Repository
Give us feedback
|Optimization Journals, Sites, Societies|