Mean robust optimization

被引:1
作者
Wang, Irina [1 ]
Becker, Cole [2 ]
Van Parys, Bart [3 ]
Stellato, Bartolomeo [1 ]
机构
[1] Princeton Univ, Dept Operat Res & Financial Engn, Princeton, NJ 08540 USA
[2] MIT, Operat Res Ctr, Cambridge, MA USA
[3] Ctr Wiskunde Informat, Stochast Grp, Amsterdam, Netherlands
关键词
Robust optimization; Distributionally robust optimization; Data-driven optimization; Machine learning; Clustering; Probabilistic guarantees; REDUCTION;
D O I
10.1007/s10107-024-02170-4
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Robust optimization is a tractable and expressive technique for decision-making under uncertainty, but it can lead to overly conservative decisions when pessimistic assumptions are made on the uncertain parameters. Wasserstein distributionally robust optimization can reduce conservatism by being data-driven, but it often leads to very large problems with prohibitive solution times. We introduce mean robust optimization, a general framework that combines the best of both worlds by providing a trade-off between computational effort and conservatism. We propose uncertainty sets constructed based on clustered data rather than on observed data points directly thereby significantly reducing problem size. By varying the number of clusters, our method bridges between robust and Wasserstein distributionally robust optimization. We show finite-sample performance guarantees and explicitly control the potential additional pessimism introduced by any clustering procedure. In addition, we prove conditions for which, when the uncertainty enters linearly in the constraints, clustering does not affect the optimal solution. We illustrate the efficiency and performance preservation of our method on several numerical examples, obtaining multiple orders of magnitude speedups in solution time with little-to-no effect on the solution quality.
引用
收藏
页数:43
相关论文
共 52 条
[1]   Tractable stochastic analysis in high dimensions via robust optimization [J].
Bandi, Chaithanya ;
Bertsimas, Dimitris .
MATHEMATICAL PROGRAMMING, 2012, 134 (01) :23-70
[2]  
Beck A., 2017, MOS-SIAM SER OPTIMIZ, DOI 10.11371.9781611974997
[3]   Robust solutions of Linear Programming problems contaminated with uncertain data [J].
Ben-Tal, A ;
Nemirovski, A .
MATHEMATICAL PROGRAMMING, 2000, 88 (03) :411-424
[4]   Selected topics in robust convex optimization [J].
Ben-Tal, Aharon ;
Nemirovski, Arkadi .
MATHEMATICAL PROGRAMMING, 2008, 112 (01) :125-158
[5]   Deriving robust counterparts of nonlinear uncertain inequalities [J].
Ben-Tal, Aharon ;
den Hertog, Dick ;
Vial, Jean-Philippe .
MATHEMATICAL PROGRAMMING, 2015, 149 (1-2) :265-299
[6]  
Ben-Tal A, 2009, PRINC SER APPL MATH, P3
[7]   A clustering approach for scenario tree reduction: an application to a stochastic programming portfolio optimization problem [J].
Beraldi, Patrizia ;
Bruni, Maria Elena .
TOP, 2014, 22 (03) :934-949
[8]   The price of robustness [J].
Bertsimas, D ;
Sim, M .
OPERATIONS RESEARCH, 2004, 52 (01) :35-53
[9]  
Bertsimas D., 2022, ROBUST ADAPTIVE OPTI
[10]   A Data-Driven Approach to Multistage Stochastic Linear Optimization [J].
Bertsimas, Dimitris ;
Shtern, Shimrit ;
Sturt, Bradley .
MANAGEMENT SCIENCE, 2023, 69 (01) :51-74