分布鲁棒二阶随机占优约束优化问题的稳定性

2023-04-08 16:15傅璐赵勇
关键词:收敛性

傅璐 赵勇

摘要:

为规避真实分布的不确定性导致的风险,提出分布鲁棒二阶随机占优约束优化问题,矩信息和Wasserstein球相结合构造分布集合,采用离散近似方法处理该问题,并在适当的假设条件下,讨论近似问题可行集、最优值和最优解集的收敛性。

关键词:二阶随机占优约束;分布鲁棒优化;收敛性

中图分类号:O224

文献标志码:A

二阶随机占优是决策论和经济学中的基本概念,作为一种稳健的风险度量,理论上可以描述任何不确定的或随机事件之间的优劣,做量化比较。在2003年,Dentcheva等[1]将二阶随机占优作为约束条件引入优化问题,研究了最优性条件和对偶理論。自此以后,许多学者对于二阶随机占优约束优化问题在最优性条件、灵敏度分析、算法等方面进行了大量研究[2-7]。但在很多实际问题中,决策者很难知道随机变量真实分布的全部信息,因此,提出分布鲁棒二阶随机占优约束优化问题,规避了真实分布的不确定性所带来的风险[8-9]。现有文献多基于矩信息定义的分布集合或者基于以经验分布为中心的Wasserstein球定义的分布集合,研究分布鲁棒二阶随机占优约束优化问题的稳定性。由于分布集合的构造对研究分布鲁棒优化问题非常重要,基于矩信息定义的分布集合不能刻画分布集合的收敛性[10],基于以经验分布为中心的Wasserstein球定义的分布集合克服了这个困难,却又失去了满足矩信息刻画的分布[11]。在此基础上,有研究改进了分布集合,提出将矩信息和Wasserstein球相结合构造分布集合,该集合不仅包含了更多的真实分布信息,又有效地排除病态分布[12]。目前,没有文献基于该分布集合研究分布鲁棒二阶随机占优约束优化问题的稳定性。参考相关文献[12],本文将基于矩信息和Wasserstein球相结合的分布集合,在适当的假设条件下,研究分布鲁棒二阶随机占优约束优化问题的稳定性。

4 结论

本文基于矩信息和Wasserstein球相结合的分布集合,在适当的假设条件下,从定性、定量的角度讨论了分布鲁棒二阶随机占优约束优化问题可行集、最优值和最优解集的收敛性。后续将基于该分布集合,研究分布鲁棒二阶随机占优约束优化问题的可处理形式,并将其应用到投资组合、供应链管理等实际问题。

参考文献

[1]DENTCHEVA D, RUSZCZYN'SKI A. Optimization with stochastic dominance constraints[J]. SIAM Journal on Optimization, 2003, 14(2): 548-566.

[2]LUEDTKE J. New  formulations for optimization under stochastic dominance constraints[J]. SIAM Journal on Optimization, 2008, 19(3): 1433-1450.

[3]FBIN C I, MITRA G, ROMAN D. Processing second-order stochastic dominance models using cutting-plane representations[J]. Mathematical Programming, 2011, 130(1): 33-57.

[4]SUN H L, XU H F, MESKARIAN R,et al. Exact penalization, level function method, and modified cutting-plane method for stochastic programs with second order stochastic dominance constraints[J]. SIAM Journal on Optimization, 2013, 23(1): 602-631.

[5]张宏伟, 于海生, 庞丽萍, 等. 二阶随机占优约束优化问题的遗传算法求解[J]. 大连理工大学学报, 2016, 56(3): 299-303.

[6]杨柳, 申飞飞. 考虑交易费用的二阶随机占优投资组合风险控制模型[J]. 应用概率统计, 2017, 33(2): 111-124.

[7]付永彬, 孙海琳. 二阶随机占优约束下考虑订购能力的多产品报童问题[J]. 工程数学学报, 2020, 37(1): 1-15.

[8]GUO S Y, XU H F,ZHANG L W. Probability approximation schemes for stochastic programs with distributionally robust second-order dominance constraints[J]. Optimization Methods and Software, 2017, 32(4): 770-789.

[9]MEI Y, LIU J, CHEN Z P. Distributionally robust second-order stochastic dominance constrained optimization with Wasserstein ball[J]. SIAM Journal on Optimization, 2022, 32(2): 715-738.

[10] SUN H L, XU H F. Convergence analysis for distributional robust optimization and equilibrium problems[J]. Mathematics of Operations Research, 2016, 41(2): 377-401.

[11] ESFAHANI P M, KUHN D. Data-driven distributionally robust optimization using the Wasserstein metric: Performance guarantees and tractable reformulations[J]. Mathematical Programming, 2018, 171(1-2): 115-166.

[12] LIU Y C, YUAN X M, ZHANG J. Discrete approximation scheme in distributionally robust optimization[J]. Numerical Mathematics: Theory, Methods and Applications, 2021, 14(2): 285-320.

[13] JIANG R W, GUAN Y P. Data-driven chance constrained stochastic program[J]. Mathematical Programming, 2016, 158(1-2): 291-327.

[14] DELAGE K, YE Y Y. Distributionally robust optimization under moment uncertainty with application to data-driven problems[J]. Operations Research, 2010, 58(3): 595-612.

[15] LIU Y C, MESKARIAN R, XU H F. Distributionally robust reward-risk ratio optimization with moment constraints[J]. SIAM Journal on Optimization, 2017, 27(2): 957-985.

[16] LIU J, LISSER A, CHEN Z P. Distributionally robust chance constrained geometric optimization[J]. Mathematics of Operations Research, 2022, 47(4): 2950-2988.

[17] ZHAO Y, LIU Y C, YANG X M. Distributionally robust reward-risk ratio programming with Wasserstein metric[J]. Pacific Journal of Optimization, 2019, 15(1): 69-90.

[18] DENTCHEVA D, RUSZCZYN'SKI A. Robust stochastic dominance and its application to risk-averse optimization[J]. Mathematical Programming, 2010, 123(1): 85-100.

[19] RMISCH W. Stability of stochastic programming problems[J]. Handbooks in Operations Research and Management Science, 2003, 10: 483-554.

[20] DUPACV J, GRWE-KUSKA N, RMISCH W. Scenario reduction in stochastic programming[J]. Mathematical Programming, 2003, 95(3): 493-511.

[21] ROBINSON S M. An application of error bounds for convex programming in a linear space[J]. SIAM Journal on Control, 1975, 13(2): 271-273.

[22] ROCKAFELLAR R T, WETS R J B. Variational analysis[M]. New York: Springer, 1998.

Stability of Distributionally Robust Second Order Stochastic Dominance Constrained Optimization Problem

FU Lu, ZHAO Yong

(College of Mathematics and Statistics, Chongqing Jiaotong University, Chongqing 400074, China)

Abstract:

In order to hedge the risks caused by the uncertainty of the true distribution, distributionally robust second order stochastic dominance constrained optimization problem was proposed, in which the ambiguity set was constructed based on the combination of moment information and Wasserstein ball. Then, the discrete approximation method was used to deal with the problem. Under appropriate assumptions, the convergence of the feasible set, the optimal value and the optimal solution set of the approximation problem was discussed.

Keywords:

second order stochastic dominance constrained; distributionally robust optimization; convergence

收稿日期:2023-06-27

基金项目:

重慶市自然科学基金面上项目(批准号:CSTB2022NSCQ-MSX1318)资助。

通信作者:

赵勇,男,博士,副教授,主要研究方向为随机优化。E-mail: zhaoyongty@126.com

猜你喜欢
收敛性
带弱阻尼Navier-Stokes方程拉回吸引子的收敛性
群体博弈的逼近定理及通有收敛性
行间AANA随机变量阵列加权和的完全矩收敛性
Lp-混合阵列的Lr收敛性
WOD随机变量序列的完全收敛性和矩完全收敛性
END随机变量序列Sung型加权和的矩完全收敛性
END随机变量序列Sung型加权和的矩完全收敛性
END序列加权和的完全收敛性
随机Kuramoto-Sivashinsky方程数值解的收敛性
行为ND随机变量阵列加权和的完全收敛性