摘要
本文考虑离散型随机向量在概率分布不完全已知情形时的二阶段问题。文中假设仅知道随机向量取有限不同值时的概率所组成的概率向量位于确定的凸多面体内。采用极大极小决策原则,不完全信息二阶段问题转化成为极大极小规划问题。本文给出求解这类极大极小规划问题在有限步内收敛的算法。
A class of two stage problems with partial information of the discrete random variables is studied in this paper. Here it is supposed that the vector, whose elements equal to the probability with which the random variables appear at some fixed values, belongs to a determined convex polyhedron. The two stage problem can be transformed to a minimax problem by means of the minimax decision rule. An efficient algorithm is designed to solve the minimax problem.
出处
《系统工程学报》
CSCD
1993年第1期102-111,共10页
Journal of Systems Engineering