An upper bound on the expected value of a non-increasing convex function with convex marginal return functions

被引：3
|
作者：
Donohue, CJ
Birge, JR
机构：
[1] Dept. of Indust. and Operations Eng., University of Michigan, Ann Arbor
来源：
OPERATIONS RESEARCH LETTERS | 1996年 / 18卷 / 05期
基金：
美国国家科学基金会;
关键词：
upper bound; convex function; approximation for stochastic programming; RECOURSE; PROGRAMS;
D O I：
10.1016/0167-6377(95)00055-0
中图分类号：
C93 [管理学]; O22 [运筹学];
学科分类号：
070105 ; 12 ; 1201 ; 1202 ; 120202 ;
摘要：
In this note, we show that if a convex function is non-increasing and has a special property we call convex marginal return functions, an effective upper bound can be established using only two function evaluations. Further, we show that this bound can be refined in such a way that the number of function evaluations needed grows linearly with the number of refinements performed.
引用
收藏
页码：213 / 221
页数：9
相关论文
共 26 条
123