移动自组网可信策略及应用结合研究

发布时间：2018-09-03 13:44

【摘要】：移动自组织网络(Mobile Ad Hoc Network,MANET)是一种由多个具有对等关系的移动无线节点组成的自组织网络,该网络旨在不依赖于任何基础设施的前提下提供网络无线服务。网络中的节点可以充任网络路由器的角色,数据包经由网络中多跳中间节点转发而到达其指定的目的节点。鉴于该网络这种自组织和快速反应部署的能力,MANET网络可以被应用到不同的场合,如战场通信、紧急营救、法庭审讯、公开会议、虚拟教室和其他一些较为安全、敏感的计算场合中。该网络因其潜在的应用远景吸引了众多的研究团队。随着移动自组网对可靠性和安全性的依赖不断增强,可信计算逐步进入到了人们的研究视野。目前,海内外就可信计算探究中已经有了很大的突破,然而移动自组网在设计可信协议、建立可信软件方面只是处于初步阶段。这当中非常有代表的就是移动自组网的安全路由协议,因为移动自组网其本身具备着开放性以及移动平台不是很可靠,所以有关恶意攻击现象还一直存在,例如自私节点攻击以及隧道攻击等。由此,本文对移动自组网下的可信策略及其相关技术进行研究：首先,综合考虑对移动自组网下影响信任关系的多种可能要素,引入模糊反馈评价理论对移动自组网的节点信息进行处理,重点剖析处理单个节点的可信评估问题。根据节点属性权重信息的不确定性和属性加权综合值划分的模糊性,并给出了相关分析和度量算法。其次,结合灰色理论和马尔可夫理论的特点,根据移动自组网中节点动态性、临时性的特性,对节点中的信任利用模型进行预测评估,结合灰色理论和马尔可夫理论的特点,用灰色SCGM(1,1)模型预测曲线来反映网络宏观发展规律,并以此预测曲线为基准,利用马尔可夫概率矩阵来寻找自组网的微观波动规律,因而对随机波动较大的数据列的预测具有较高的精度预测值与实际值较为接近,较好的反应了节点的变化关系,所得到的预测值是可信的。同时依据节点的信任预测提出了一种新型的可信源路由协议。在移动自组网中,由于节点设备能量有限,如何构造最小能耗多播树是一个重要问题。由于构造最小能耗多播树受到选择不同的中继节点集的影响,论文提出了一种优化最小能耗多播树构造的可信离散粒子群算法。实验结果表明改进后的可信离散粒子群算法具有较强的优化能力,有效地提高了最小能耗多播树的构造。最后,针对移动自组网中的信任机制和调度算法相分离,从而在实际环境中难以应用的缺陷,论文引入了多目标遗传算法来刻画网络任务调度的多目标遗传模型,提出了一种基于信任机制与改进的遗传任务调度算法,并对算法进行改进以保证算法的收敛性。
[Abstract]:Mobile Ad Hoc Network (Mobile Ad Hoc Network,MANET) is an ad hoc network composed of multiple peer-to-peer mobile wireless nodes, which is designed to provide network wireless services without relying on any infrastructure. The nodes in the network can act as network routers, and the packets are forwarded to the designated destination nodes by the multi-hop intermediate nodes in the network. In view of the network's self-organizing and rapid response deployment capability, Manet networks can be applied to different situations, such as battlefield communications, emergency rescue, court hearings, public meetings, virtual classrooms and others, In sensitive computing situations. The network has attracted a large number of research teams for its potential application prospects. With the increasing dependence of Manet on reliability and security, trusted computing has gradually entered the field of study. At present, there has been a great breakthrough in the research of trusted computing at home and abroad. However, the design of trusted protocols and the establishment of trusted software in Manet are only in the initial stage. This is very representative of the mobile ad hoc network security routing protocol, because the mobile ad hoc network itself is open and mobile platform is not very reliable, so malicious attacks have always existed. Such as selfish node attack and tunnel attack. Therefore, this paper studies the trusted strategy and its related technologies in Manet. Firstly, we consider a variety of possible factors that affect the trust relationship under Manet. The fuzzy feedback evaluation theory is introduced to deal with the node information of mobile ad hoc network, and the trust evaluation problem of a single node is analyzed and dealt with. According to the uncertainty of attribute weight information and the fuzziness of attribute weighted comprehensive value partition, the correlation analysis and measurement algorithm are given. Secondly, combining the characteristics of grey theory and Markov theory, according to the dynamic and temporary characteristics of nodes in mobile ad hoc networks, the trust utilization model in nodes is forecasted and evaluated, and the characteristics of grey theory and Markov theory are combined. The prediction curve of grey SCGM (1K1) model is used to reflect the macroscopic development law of the network. Based on the prediction curve, the microscopic fluctuation law of the ad hoc network is found by using Markov probability matrix. Therefore, the prediction of the random fluctuating data series has higher accuracy and close to the actual value, which reflects the change of the node, and the predicted value is credible. At the same time, based on the trust prediction of nodes, a new source routing protocol is proposed. Due to the limited energy of node devices in mobile ad hoc networks, how to construct the minimum energy consumption multicast tree is an important problem. Because the minimum energy consumption multicast tree is affected by the selection of different relay node sets, a trusted discrete particle swarm optimization algorithm is proposed to optimize the construction of the minimum energy consumption multicast tree. Experimental results show that the improved trusted discrete particle swarm optimization algorithm has a strong ability to optimize and effectively improve the construction of the minimum energy consumption multicast tree. Finally, aiming at the shortcomings of the separation of trust mechanism and scheduling algorithm in Manet, which is difficult to apply in real environment, this paper introduces a multi-objective genetic algorithm to describe the multi-objective genetic model of network task scheduling. A genetic task scheduling algorithm based on trust mechanism and improved genetic algorithm is proposed and improved to ensure the convergence of the algorithm.
【学位授予单位】：山东大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN929.5

【相似文献】