摘要
针对分布鲁棒问题的保守性,利用凸分析中的理论研究了一类混合离散分布信息下的全局分布鲁棒问题的等价形式.当只有概率分布是不确定变量时,得到了相应全局分布鲁棒优化问题的易计算的确定问题;当样本值与概率分布均不确定时,得到其全局分布鲁棒优化问题的等价确定形式.
For the conservation of the solution of robust optimization problem,we consider a class of globalized distributionally robust optimization problem under mixture discrete distribution infformation by the convex analysis.When probability vector is uncertain,we get the explicit and computationally tractable problem for globalized distributionally robust optimization problem.When samples and their probabilities are both uncertain,we present an explicit problem for globalized distributionally robust optimization problem.
出处
《四川师范大学学报(自然科学版)》
CAS
北大核心
2017年第3期334-339,共6页
Journal of Sichuan Normal University(Natural Science)
基金
四川省教育厅科研基金(16ZB0080)
中央高校基本科研业务费专项项目(2014NZYQN49)
关键词
全局鲁棒问题
分布鲁棒
共轭函数
支撑函数
globalized robust problem
distributionally robust
conjugate functions
support functions