STOCHASTIC SADDLE POINT PROBLEMS WITH DECISION-DEPENDENT DISTRIBUTIONS

被引:6
作者
Wood, Killian [1 ]
Dall'Anese, Emiliano [1 ,2 ]
机构
[1] Univ Colorado Boulder, Dept Appl Math, Boulder, CO 80309 USA
[2] Univ Colorado Boulder, Dept Elect Comp & Energy Engn, Boulder, CO USA
基金
美国国家科学基金会;
关键词
saddle point problems; minimax problems; stochastic gradient; decision-dependent distributions; OPTIMIZATION; ALGORITHMS; COMMUNICATION; GRADIENT;
D O I
10.1137/22M1488077
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
This paper focuses on stochastic saddle point problems with decision-dependent distributions. These are problems whose objective is the expected value of a stochastic payoff function and whose data distribution drifts in response to decision variables-a phenomenon represented by a distributional map. A common approach to accommodating distributional shift is to retrain optimal decisions once a new distribution is revealed, or repeated retraining. We introduce the notion of equilibrium points, which are the fixed points of this repeated retraining procedure, and provide sufficient conditions for their existence and uniqueness. To find equilibrium points, we develop deterministic and stochastic primal-dual algorithms and demonstrate their convergence with constant step size in the former and polynomial decay step-size schedule in the latter. By modeling errors emerging from a stochastic gradient estimator as sub-Weibull random variables, we provide error bounds in expectation and in high probability that hold for each iteration. Without additional knowledge of the distributional map, computing saddle points is intractable. Thus we propose a condition on the distributional map-which we call opposing mixture dominance-that ensures that the objective is strongly-convex-strongly-concave. Finally, we demonstrate that derivative-free algorithms with a single function evaluation are capable of approximating saddle points.
引用
收藏
页码:1943 / 1967
页数:25
相关论文
共 50 条
[31]   Sparse Polynomial Approximations for Affine Parametric Saddle Point Problems [J].
Chen, Peng ;
Ghattas, Omar .
VIETNAM JOURNAL OF MATHEMATICS, 2023, 51 (01) :151-175
[32]   Data-driven project portfolio selection: Decision-dependent stochastic programming formulations with reliability and time to market requirements [J].
Kettunen, Janne ;
Lejeune, Miguel A. .
COMPUTERS & OPERATIONS RESEARCH, 2022, 143
[33]   Decision-Dependent Multiobjective Multiperiod Stochastic Model for Parking Location Analysis in Sustainable Cities: Evidence from a Real Case [J].
Aydin, Nezir .
JOURNAL OF URBAN PLANNING AND DEVELOPMENT, 2022, 148 (01)
[34]   Decision-dependent distributionally robust Markov decision process method in dynamic epidemic control [J].
Song, Jun ;
Yang, William ;
Zhao, Chaoyue .
IISE TRANSACTIONS, 2024, 56 (04) :458-470
[35]   Decision-Dependent Uncertainty Modeling in Power System Operational Reliability Evaluations [J].
Hu, Bo ;
Pan, Congcong ;
Shao, Changzheng ;
Xie, Kaigui ;
Niu, Tao ;
Li, Chunyan ;
Peng, Lvbin .
IEEE TRANSACTIONS ON POWER SYSTEMS, 2021, 36 (06) :5708-5721
[36]   Robust approximation of chance constrained DC optimal power flow under decision-dependent uncertainty [J].
Aigner, Kevin-Martin ;
Clarner, Jan-Patrick ;
Liers, Frauke ;
Martin, Alexander .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2022, 301 (01) :318-333
[37]   Modified Uzawa methods for saddle point problems [J].
Li, Chen ;
Shao, Xin-Hui ;
Li, Chang-Jun .
APPLIED MATHEMATICS AND COMPUTATION, 2017, 315 :507-515
[38]   A relaxed splitting preconditioner for saddle point problems [J].
Li, Cui-Xia ;
Wu, Shi-Liang ;
Huang, Ai-Qun .
JOURNAL OF NUMERICAL MATHEMATICS, 2015, 23 (04) :361-368
[39]   Adaptive wavelet methods for saddle point problems [J].
Dahlke, S ;
Hochmuth, R ;
Urban, K .
ESAIM-MATHEMATICAL MODELLING AND NUMERICAL ANALYSIS-MODELISATION MATHEMATIQUE ET ANALYSE NUMERIQUE, 2000, 34 (05) :1003-1022
[40]   On parameter acceleration methods for saddle point problems [J].
Zhang, Naimin .
JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS, 2015, 288 :169-179