纳什均衡策略的极大熵估计方法

给出了策略熵的定义,论证了最大策略熵是纳什均衡的充分必要条件,并探讨了运用熵极大化准则来估计博弈参与者的混合策略的方法,得到了与传统方法相同的纳什均衡解,这表明纳什均衡策略是既定的收益约束条件的最大熵,为纳什均衡提供了基于信息论的解释,同时也为求解纳什均衡提供了一种极大熵估计方法,算例验证了该方法的可行性和有效性。

Estimating Mixed Nash Equilibrium Employing Maximum Entropy Inference

We presented the notion of strategic entropy at the beginning of this paper and discussed the relationship between the maximum strategic entropy and Nash Equilibria. Then we employed the maximum entropy inference to estimate mixed Nash equilibrium and got the same results as classical methods. And we demonstrated that strategic entropy under Nash Equilibrium is the maximum entropy, which produced an information-theoretic explanation of game theory and a new method to calculate mixed Nash equilibrium. A practical example ended our paper to confirm the feasibility and validity.