【作者】 牛新征；

【导师】 周明天；

【作者基本信息】 电子科技大学 ， 计算机应用技术， 2008， 博士

【摘要】 人们对通过个人数字助理等设备实现资源的协作和共享等移动性的应用需求，为移动对等网络(即移动P2P网络)的学术研究和技术开发带来了新的机遇。本文系统地概述了移动P2P网络技术，包括研究背景、研究现状和关键技术研究分类，并从移动P2P网络的体系结构、协作激励机制、资源共享方案等关键技术的研究入手，深入研究了移动P2P网络计算技术。研究工作取得了如下创新性成果：(1)提出一种新的移动P2P路由策略路由策略是资源共享方案研究中最重要的部分。本文提出了一种新的采用有限洪泛路由查询和移动Agent路由查询相结合的路由发现策略。同时，提出了新的基于热度的蚁群寻路式路由选择算法，和综合考虑多个路由性能指标和解决路由错误的路由保持机制。本文提出的策略在控制消息的开销等方面具有优良性能。(2)路由信息的模糊知识处理算法基于上述路由策略，本文提出了一种从移动Agent所携带的模糊知识中分析和过滤出较准确和有价值路由数据的算法。该算法与目前的基于移动Agent的移动ad hoc网络路由算法相比，有效减小了平均端到端时延和路由开销等。对于因节点的移动等变化而引起的路由信息变化性、模糊性有较好的适应度。(3)提出了一种基于博弈论的协作激励机制针对移动节点的资源有限和部分节点具有自私行为的特点，本文提出了一种基于博弈论的协作激励机制。文中还详细描述和分析了协作激励机制中的网络资源分配策略，证明了该协作激励机制的博弈存在一个稳定的纳什均衡。该激励机制有效地激发了移动节点间的协作，提高了数据包转发率等。(4)移动节点间协作资源请求调度方案本文提出的请求调度方案既为协作资源的请求者设计了基于可靠性理论的协作资源请求预测算法，也为协作资源的提供者构建了基于排队论的层次型协作资源的请求执行调度模型。理论分析证明，预测算法提高了节点协作资源请求任务的执行成功率，层次型资源调度模型则提高了节点的协作共享资源的服务能力等。(5)提出了一种协作共享缓存-Cache资源即协作资源的优化策略为了合理高效地使用移动节点提供的协作共享缓存-Cache资源本身，本文提出了一种基于信息素的协作Cache资源替换算法，建立了选取适当存储空间作为协作Cache资源的数学模型和推导。该算法能有效提高协作Cache资源的使用率，并提高了移动节点从协作节点获得协作缓存-Cache资源的成功率等。论文根据可计算运输系统项目，进行了体系结构的设计以及实现了一种移动P2P网络实际应用的原型系统，并给出了关键模块和技术的具体设计等。更多还原

【Abstract】 There has the new application requirement for achieving mobile users’s resources collaborating and sharing tasks through different mobile devices such as personal digital assistant and so on. So these bring a new demand for the academic research and industrial development of mobile P2P networks. This dissertation introduces the technology of mobile P2P networks from a full-scale vision, such as research background, present research situation and the classification of research work on key issues, etc. According to the classification of research work on key issues, the technology of mobile P2P networks is deep studied through the design of architecture, nodes cooperative incentive mechanism, resource sharing scheme and so on. Some innovative contributions of the dissertation are enumerated as follows:(1) Proposing a new mobile P2P routing algorithmRouting algorithm is the most important part of resource sharing scheme research. The dissertation proposes a routing discovery algorithm based on the restriction flooding and mobile Agents routing search. Moreover, in this dissertation the new improved ant colony routing selection algorithm is used based on the heat degree. Routing performance objects and methods resolving bursty routing problems are synthetically analyzed for routing maintenance algorithm. This algorithm performs well in controlling the route overhead, etc.(2) Proposing a fuzzy knowledge processing algorithm for route informationAn algorithm which can analyze and filter the valuable data from the fuzzy information carried by mobile Agents is proposed based on the routing algorithm mentioned above. Compared with the existing routing algorithms in mobile ad hoc networks which are based on mobile Agents and the routing algorithm, the results of this dissertation show the proposed algorithm can reduce the average end-to-end delay and routing overhead, and so on. It also can have the better adaptability to the variability and fuzziness of routing information which is caused by the mobility of mobile peers.(3) Proposing a cooperative incentive mechanism based on game theoryConsidering facing to the resource limitation and selfishness of mobile peers, this dissertation proposes a cooperative incentive mechanisms based on game theory. This dissertation describes and analyzes the resource allocation policy of cooperative incentive mechanism in detail. Besides, the existence of the steady Nash equilibrium in the game is also proven. The incentive mechanism can encourage the cooperation among the peers and increase forwarding ratio of the packet and so on.(4) Proposing a resource requesting and dispatching scheme among mobile peersThe resource requesting and dispatching scheme among mobile nodes design aprediction algorithm of resources based on reliability theory for resources’ requestor. A multi-level model of resource dispatching has been put forward for resources’ provider based on the queue theory. According to theory analyse and derivation, the prediction algorithm can increase success ratio of sharing resources’ cooperation task and the resources dispatching model can significantly improve the service quality and so on.(5) Proposing an optimized policy of cooperative sharing Cache resourcesIn order to make full use mobile nodes’ cooperative Cache resources effectively, a cooperative Cache resources replacement policy based on the pheromone is proposed, and a mathematical model and the derivation of choosing appropriate memory space as cooperative Cache resources is established. The new policy can significantly improve cooperative Cache resource utility. It also can promote the success ratio of resource obtainment from cooperation nodes and so on.The architecture of mobile P2P networks is designed and the prototype system of a practical mobile P2P application is implemented based on the project of computational transportation system. The design details of some of the key modules and issues are also provided in this dissertation.更多还原