基于序数信息的博弈理论

在许多社会或政治科学的问题中, 个人的主观偏好信息往往是通过定性分析得到的, 难 以表示成效用形式, 但却可以用序数形式来表示. 本文基于决策者对策略的偏好序数信息而不 是支付函数或效用水平, 讨论了一种新的博弈理论或模型, 给出了若干序均衡的概念, 并进行 了例示分析. 这种基于序数信息的博弈理论可以看作是现有基于支付函数博弈理论的推广和 补充.

p reference rank o rder; gam e theo ry; N ash equ ilib rium; Pareto equ ilib rium; Stackelberg equ ilib rium

In m any social o r po lit ical science p rob lem s, the personal sub ject ive p reference info rm a2 t ion can be the resu lt f rom qualitat ive analysis, w h ich is diff icu lt to be exp ressed in the u t ility fo rm , bu t can be exp ressed in o rdinal fo rm. Based upon o rdinal p reference info rm at ion the decision2m akers exp ressed, in stead of the payoff funct ion o r u t ility level, th is paper developed a rank2o rdering based gam e theo ry o r model, p ropo sed som e concep t s of rank2o rdering equ ilib rium , and gave som e num eri2 cal demon st rat ion s. Th is new gam e theo ry cou ld be con sidered as the necessary and valuab le general2 izat ion and supp lem en t to the t radit ional u t ility o r payoff based gam e theo ry.