这个跟prisoner dilemma很相似,NE {5000,5000} 是inefficient.
{20.000,20.000}不是NE因为有一个incentive to deviate
但是是Pareto dominant{5000,5000}
Rationality有很多定义,但没有一个是说outcome必须是efficient.
比如 rational agents behave as if they only care about monetary payoff.
Q1:参与人: player 1, player 2
参与人可选的策略: any umber [5000,20.000]
PS: interva l[5000, 20.000]是continue的,所以每个人都有无数的actions
收益: Min {A.B}
player 1: if A>B, then A'=B-500 Player 2: B'=B+500
A<B,then A'=A+500 B'=A-500