The Application Of Evolutionary Games With Incomplete Information On The Endowment Insurance Of Pay Cost

In the classical game theory, participants is always assumed to be perfectly rational and selfish.They usually have enough analysis ability. But in real life, this assumption is almost impossible. So people put forward the evolutionary game theory,turnning the static analysis of traditional game into a dynamic process of constant evolution.In the dynamic evolution process, participants learn from the others to get more profits, in accordance with the traditional strategy update rules, which requires participants to master the opponent's earnings information. But in real life, the participants are unable to get all the information of opponents in most cases.At first,the thesis divide the traditional behavior strategy space into four areas,because of the current widely used strategy update rules can't be used in incomplete information game theory ,the thesis proposed a new strategy update rules which based on incomplete information and emphasize the influence of historical returns to the current behavior choice.The thesis analyze the evolution of the prisoner's dilemma based on single network configurations such as BA scale-free network, WS small-world network,ER random graph and heterogeneous networks, to compare the impact of different network configuration, network scale and the average degree on gambling behavior.Finally the thesis apply this strategy update rules to endowment insurance expends game,to analyze the influence of different level of wages, punish multiples,selective examination proportion and verification accuracy on endowment insurance pay cost behavior.