摘要
借鉴进化博弈的思想,一个新的模型建立起来用以模拟造假行为的传播过程,以期找到其广泛传播的路径,并寻找遏制的方法。模型赋予了个体学习的能力,结果显示惩罚程度对造假行为的传播没有影响,但抽查概率却是决定性的。为了提高查处的效率,"连带检查"机制被加入到模型中,即对造假者周围邻居进行检查并施以同样的惩罚。这种机制能够在抽取概率为26%时达到消除造假者的目的,相比之下,没有"连带检查"时必须把抽取概率提高到80%才能达到消除造假者的目的。
An evolutionary game model is built up to simulate the fraud behaviors and find a way to better contain them. With modifications the classical Pavlov strategy is changed with learning ability. The simulations depict that the punishment degree doesn＇t affect much, on the spread of frand behaviors but the checking frequency is a determining factor. To improve efficiency,the ＂collateral checking＂ mechanism is adopted to check and punish the frauds as well as the frauds in their neighborhood. The new mechanism shows better result since it can eliminate frauds at the checking percentage of 26%0, while it has to be raised to 80% if without ＂collateral checking＂.
出处
《武汉大学学报（哲学社会科学版）》
CSSCI
北大核心
2014年第3期112-118,共7页
Wuhan University Journal：Philosophy & Social Science
基金
国家重点基础研究发展计划项目(2011CB302306)
武汉大学研究生自主科研项目(201121601020001)
关键词
进化博弈
造假行为
惩罚机制
连带检查
evolutionary game
fraud behavior
punishment method
collateral checking
