嵌套信息结构下的多级递阶诱导决策问题

Incentive Mechanism for Hierarchical Systems Under Nested Information Structure

本文基于2人2级stackelberg鼓励性对策下的广义平衡解的定义,引出决策可行区和可接受区的概念,推出广义平衡解存在的充分必要条件,并拓广到多人多级情形.总结和应用等价原则对多级递阶系统进行简约,在适当条件下,可将多级决策系统简化成单人决策问题.通过探讨广义平衡解的决策过程,揭示嵌套信息结构下可接受区的反嵌套关系,从而有助于理解和应用带鼓励性对策的多级递阶决策方法.

In this paper, a definition of a generalized equilibrium solution is given for a two - person - two - level incentive stackelberg game, which is extended to one -leader N - follower case and N + 1 -person N + 1 - level case. The sufficient and necessary conditions for the existence of the equilibrium solutions found out. Several equivalence principles are presented, which could be used for the reduction of hierarchical sytems into simpler cases.