一种能快速收敛的对等网络信任值计算算法

A fast convergent algorithm for computing the trust value in peer-to-peer networks

对目前对等网络中常用的信任值计算方法进行了形式化分析,提出了一个分布式信任值迭代方法,根据其特点命名为阻尼方法,理论证明阻尼方法一定收敛.在对阻尼方法收敛速度和安全性的分析基础上,提出了一个结合了名誉管理、自适应调整的快速信任值计算算法(TVCA).模拟实验表明TVCA算法在提高计算效率、抵抗恶意攻击等方面都能取得良好的效果.

In the researches of the trust management of the peer-to-peer networks, the computing of trust value is most important. Typical algorithms nowadays for computing the trust value in peerto-peer networks are formally analyzed. The results show that these algorithms cant produce the correct trust value. Therefore, a distributed iterative algorithm to compute the trust Value called as damping method by its properties is provided. The formal analyses show that damping method must be convergent. Based on the analyses of the convergent speed and security of damping method, a fast convergent trust value computing algorithm or TVCA combining reputation management and self-adaptive adjustment is provided. The experimental results show that TVCA can achieve good effects in many aspects such as efficient computing and malicious attack resistance.