SCENARIO MIN-MAX OPTIMIZATION AND THE RISK OF EMPIRICAL COSTS

被引:49
作者
Care, A. [1 ]
Garatti, S. [2 ]
Campi, M. C. [3 ]
机构
[1] Univ Melbourne, Dept Elect & Elect Engn, Parkville, Vic 3052, Australia
[2] Politecn Milan, Dipartimento Elettron Informaz & Bioingn, I-20133 Milan, Italy
[3] Univ Brescia, Dipartimento Ingn Informaz, I-25123 Brescia, Italy
关键词
stochastic optimization; sample-based techniques; scenario approach; data-driven optimization; PROBABILISTIC OPTIMIZATION; DISCRETE-DISTRIBUTIONS; RANDOMIZED SOLUTIONS; CHANCE CONSTRAINTS; PROGRAMS; STABILITY; FEASIBILITY; DESIGN;
D O I
10.1137/130928546
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
We consider convex optimization problems in the presence of stochastic uncertainty. The min-max sample-based solution is the solution obtained by minimizing the max of the cost functions corresponding to a finite sample of the uncertainty parameter. The empirical costs are instead the cost values that the solution incurs for the various parameter realizations that have been sampled. Our goal is to evaluate the risks associated with the empirical costs, where the risk associated with a cost is the probability that the cost is exceeded when a new realization of the uncertainty parameter is seen. This task is accomplished without resorting to uncertainty realizations other than those used in optimization. The theoretical result proved in this paper is that these risks form a random vector whose probability distribution is an ordered Dirichlet distribution, irrespective of the probability measure of the stochastic uncertainty parameter. This result provides a distribution-free characterization of the risks associated with the empirical costs that can be used in a variety of application problems.
引用
收藏
页码:2061 / 2080
页数:20
相关论文
共 54 条
  • [1] Randomized methods for design of uncertain systems: Sample complexity and sequential algorithms
    Alamo, Teodoro
    Tempo, Roberto
    Luque, Amalia
    Ramirez, Daniel R.
    [J]. AUTOMATICA, 2015, 52 : 160 - 172
  • [2] Randomized Strategies for Probabilistic Solutions of Uncertain Feasibility and Optimization Problems
    Alamo, Teodoro
    Tempo, Roberto
    Camacho, Eduardo F.
    [J]. IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2009, 54 (11) : 2545 - 2559
  • [3] [Anonymous], 2004, Central European Journal of Operations Research
  • [4] Practical representations of incomplete probabilistic knowledge
    Baudrit, C.
    Dubois, D.
    [J]. COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2006, 51 (01) : 86 - 108
  • [5] Bayraksan G., 2009, Tutorials in Operations Research, P102, DOI DOI 10.1287/EDUC.1090.0065
  • [6] Assessing solution quality in stochastic programs
    Bayraksan, Guezin
    Morton, David P.
    [J]. MATHEMATICAL PROGRAMMING, 2006, 108 (2-3) : 495 - 514
  • [7] Bertsimas D., 2006, TUTORIALS OPERATIONS
  • [8] Bertsimas D., 2006, Optimization Online, P1
  • [9] Uncertain convex programs: randomized solutions and confidence levels
    Calafiore, G
    Campi, MC
    [J]. MATHEMATICAL PROGRAMMING, 2005, 102 (01) : 25 - 46
  • [10] The scenario approach to robust control design
    Calafiore, Giuseppe C.
    Campi, Marco C.
    [J]. IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2006, 51 (05) : 742 - 753