ROUTING-PROBLEMS AND MARKOVIAN DECISION-PROCESSES

被引：4

作者：

SANCHO, NGF

机构：

[1] McGill Univ, Dep of Mathematics, & Statistics, Montreal, Que, Can, McGill Univ, Dep of Mathematics & Statistics, Montreal, Que, Can

来源：

JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS | 1985年 / 105卷 / 01期

关键词：

DECISION THEORY AND ANALYSIS;

D O I：

10.1016/0022-247X(85)90097-6

中图分类号：

O29 [应用数学];

学科分类号：

070104 ;

摘要：

The problem of finding a routing problem through a network has been solved by dynamic programming. A different version of the problem is now introduced, where at each city there is a transition probability p//i//j of going from city i to city j. The expected time of going from city i to N (end of destination) is then calculated. If, however, at each city there is a choice of policies (one might decide to travel from city i to city j by a different method of transport) then a new Markovian decision process is formulated. The method is easily extended for finding the maximum reliability from city i to city N using an optimal policy.

引用

页码：76 / 83

页数：8