Title :
On convergence of evolutionary games
Author :
Daizhan Cheng ; Hongsheng Qi ; Yuanhua Wang ; Ting Liu
Author_Institution :
Inst. of Control Sci. & Eng., Shandong Univ., Ji´nan, China
Abstract :
The set of finite games with fixed numbers of players and strategies for every player becomes a vector space. Certain equivalences are introduced to classify the elements in the vector space of finite games. Then the subspace of (exact or weighted) potential games are calculated. For the evolutionary (finite) games, certain strategy updating rules are investigated, which lead to certain profile dynamics consisting with the equivalence. The convergence to (pure) Nash equilibriums is investigated. Finally, the projection of finite games to the subspace of potential games is considered, and a simple formula is given to calculate the projection. The dynamics between a game and its projection is compared, which produces a method to verify the convergence of an evolutionary game to a Nash equilibrium or an ε-equilibrium.
Keywords :
evolutionary computation; game theory; Nash equilibriums; evolutionary games convergence; finite games set; strategy updating rules; Aerospace electronics; Convergence; Electronic mail; Games; Nash equilibrium; Support vector machine classification; Vectors; Nash equilibrium; Potential game; convergence; evolutionary game; sequential or cascading myopic best response adjustment rule;
Conference_Titel :
Control Conference (CCC), 2014 33rd Chinese
Conference_Location :
Nanjing
DOI :
10.1109/ChiCC.2014.6895886