求解图型博弈的Nash均衡

Finding Nash Equilibria in Graphical Game

图型博弈是一种新的博弈表示方法。求解Nash均衡是图型博弈的核心问题。论文把求解图型博弈的Nash均衡看作是离散空间中的优化问题,给出了求解图型博弈ε-Nash均衡的迭代优化算法。另外,为加快算法的收敛速度,提出了一个获得高迭代效率策略剖面的方法:基于博弈的图形结构进行多策略更新。实验结果表明算法是可行、高效的。

Graphical Game is a compact representation for mul ti-player game theory.Finding Nash Equilibria is the most important task of Gr aphical Game.We consider this task as an optimization problem in a discretized space,and present an iterative optimization algorithm for finding?-Nash Equ ilibrium profiles of Graphical Game.To make the iter-ations more effective,we provide a new method to get iterative strategy profiles based on the structure of Graphical Game.The experiment results show our method is viable and effic ient.