Wright-Fisher Multi-Strategy Trust Evolution Model of Internetware

网构软件是开放网络环境中软件系统基本形态的一种抽象,其信任关系本质上是最复杂的社会关系之一.为了增强信任演化模型的自适应性,提高预测的准确性以及有效地抑制自私节点的产生,结合差异化服务和演化博弈理论,提出了一种符合开放网络特征的信任演化模型:(1) 建立基于差异化服务的实体全局收益函数,以增强信任演化模型的自适应性;(2) 以演化博弈理论作为分析工具,借助Wright-Fisher 模型的特点,提出了一种Wright-Fisher 多策略信任演化模型,以增强对信任演化预测的准确性;(3) 根据“公平规范”原则建立了基于博弈的激励机制,激励信任策略的演化,从而有效地抑制自私节点的产生.实验结果表明:该模型能够更准确地反映开放网络中实体信任行为的复杂性特点;在激励机制的作用下,网构软件的信任演化能够更快地达到稳定状态,从而有效地提高网络的效率,使网构软件系统的信任收益达到最优.;Internetware is an abstract of software system in open network environment. Its trust relationship is one of the most complex social relationships. In order to enhance adaptability of trust evolution model, improve prediction accuracy about trust evolution, and restrain production of selfish nodes effectively through discriminatory service and evolutionary game theories, this paper puts forward a trust evolution model that corresponds with characters of an open network: (1) Global profit function of entities based on discriminatory service is gained for enhancing adaptability of trust evolution model; (2) With the help of evolutionary game theory, and based on characters of Wright-Fisher model, a kind of Wright-Fisher multi-strategy trust evolution model of internetware is proposed for enhancing the prediction accuracy about trust evolution; (3) According to principle of fairness, incentive mechanism based on game is built, so as to inspire evolution of trust strategy and restrain production of selfish nodes effectively. The experimental results show that this model can more accurately reflect complexity character of open network. By adding incentive mechanism, internetware system can achieve the stable state more quickly, thus it can improve the efficiency of network more effectively and make the trust profit achieve optimal.