欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对现有的信任模型不能很好地处理P2P网络环境中存在的不诚实推荐、协同作弊与复杂的策略性欺骗等问题，提出了一种具有激励效果的分布式P2P信任管理模型（IMTM），以量化和评估节点的可信程度，并给出了模型的数学表述和分布式实现方法。信任模型利用相似度函数构建的推荐可信度模型作为节点推荐的准确性度量，并充分考虑节点行为的稳定性、可信服务持续强度、风险性及信任的时间相关性等因素。除此之外，为了有效激励节点积极提供高质量的服务，惩罚不良行为节点，还提出一种基于节点信任度的激励机制。分析及仿真实验表明，与已有的信任模型相比，IMTM在抑制更广泛的恶意节点攻击行为中，表现出更好的适应性、灵敏性与有效性。
英文摘要：
An important challenge regarding peers＇ trust valuation in peer-to-peer （P2P） systems was how to cope with such issues as dishonest feedbacks from malicious peers, collusions and complicatedly strategic frauds to the trust model itself, which could not be effectively tackled by the existing solutions. Thus, an incentive mechanism integrated trust management model for P2P networks, （IMTM） was proposed, to quantify and evaluate the trustworthiness of peers. Moreover, the related definitions and distributed implementation strategies of IMTM was also given. In IMTM, the rec- ommendation credibility, composed of three factors, including the similarity characteristics, the altering scope of interaction experiences and the time fading feature of trust when interacting, was introduced to portray the extent to which the trustor trusted another recommender＇s recommendations. Besides, the trusted service persistent intensity was imported to this model to promote peers to provide high-quality services to others continuously, and the risk mechanism was used to improve the sensitivity of this model. Additionally, based on the trust model, an adaptive reputation based incentive mechanism was presented. Through this mechanism, trusted peers could migrate to the centric position, while untrusted peers to the edge of the topology, incenting peers to provide more high-quality services in order to get more return on services. Theoretical analyses and experimental results demonstrate that IMTM has advantages in combating such malicious behaviors as the dishonest recommendations from malicious peers, collusions and the complicatedly strategic attacks to the trust model itself over the existing models, and show more adaptability, sensitivity and effectiveness.
同期刊论文项目
同项目期刊论文
期刊信息
位置：