摘要
讨论应用随机优势原则求解随机决策问题的对话式决策方法.文中首先证明了在一定条件下可用一系列已知函数的正线性组合逼近决策人的效用函数所在的类,在此基础上,给出了在决策人提供新信息时确定新的效用函数类并在此新类中判断两方案随机优势关系的方法,进而提出了使决策人隐知的效用函数的数学期望达到最大的对话式决策程序.
This paper discusses an interactive procedure which exploits stochastic dominance rule solve stochastic decision making problems.It is proved that a decision maker's utility function class can be appropriated by a set consisted of all positive linear combinations of a series of known functions under certain conditions,then,some results are obtained for determining the new utility function class after decision maker's expressing his further imformation about the decision making problem and for judging the stochastic dominant relation between alternatives,finally,an interactive procedure is proposed to determine the action which maximizes the expectation of decision maker's implied utility function.
出处
《系统工程学报》
CSCD
1998年第1期21-28,共8页
Journal of Systems Engineering
关键词
随机决策
效用函数
随机优势
数学期望
对话式决策
stochastic decision making,utility function,stochastic dominance,interactive procedure of decision making