Optimality Condition and Complexity of Order-Value Optimization Problems and Low Order-Value Optimization Problems

Zhongyi Jiang,Qiying Hu,Xiaojin Zheng
DOI: https://doi.org/10.1007/s10898-017-0520-2
2017-01-01
Journal of Global Optimization
Abstract:Order-value optimization problems and low order-value optimization problems are new subclasses of optimization problems which arise from decision-making problems under uncertainty and robust estimation problems. In this paper, We present KKT necessary and sufficient conditions for low order-value optimization problems under convexity hypothesis in this paper. A smooth reformulation of low order-value optimization problems are presented whose local solutions satisfy the KKT necessary conditions. we prove that order-value optimization problems is NP-hard in the strong sense even when constraints are polytope. Order-value optimization problems and low order-value optimization problems are NP-hard even when their presentation functions are linear and constraints are polytope. Some special cases that could be solved in polynomial time are proposed.
What problem does this paper attempt to address?