摘要
针对带有矩约束的两阶段分布式鲁棒优化问题,当随机变量的支撑集是多面体时,利用线性规划对偶、无穷维规划对偶、二次规划的Wolfe对偶等理论研究两阶段分布式鲁棒优化问题的等价可求解模型.在分布式鲁棒优化的决策变量服从线性决策和第二阶段中的右端项为随机变量两种不同的情形下,给出对应的两阶段分布式鲁棒优化均能等价转化为可用已有算法求解的二阶锥优化问题.
For a two-stage distributionally robust optimization with moment constraints,when the supporting set of the random variable is polyhedral set, the equivalent solvable model of two-stage distributionally robust optimization problem is studied by using the dual theory of linear programming, the dual of infinite-dimensional programming, and the Wolfe dual of quadratic programming. When the decision variables of distributionally robust optimization problems obey linear decision rule and the right-hand side in the second stage is only random variable,we obtain the conclusions that the corresponding two-stage distributionally robust optimization is equivalent to the corresponding second-order cone optimization problem that can be solved by existing algorithms.
作者
张轩
韩有攀
ZHANG Xuan;HAN Youpan(School of Science,Xi′an Polytechnic University,Xi′an 710048,China)
出处
《西安工程大学学报》
CAS
2018年第5期616-621,共6页
Journal of Xi’an Polytechnic University
基金
国家自然科学基金青年科学基金(11501434)
关键词
分布式鲁棒优化
矩约束
二阶锥优化
线性决策
Wolfe对偶
Distributionally robust optimization
moment constraints
second-order cone optimization
linear decision rule
Wolfe dual
