车载网中路由算法研究

发布时间：2018-03-17 08:03

本文选题：车载网　切入点：群算法　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：近几年,随着云计算,大数据等新兴技术的发展,智慧城市、智能交通等概念也走进人们的视野。车载网以其独特的魅力和广阔的市场前景吸引人们的目光。因此,针对车载网路由算法的研究也成为了一个热点。本文结合车载网自身的特点,分析了Q-Learning,蚁群算法和模糊逻辑推理,以及目前存在的一些路由算法。针对一些算法的缺点,得到改进方法,并通过仿真实验证明,改进后的算法的可行性。本文研究了Q-Learning和蚁群算法在车载网路由算法中的应用。分析Q-ABR算法,指出其信息素更新方式的不足之处,在此基础上提出信息素更新的改进方法。针对算法设计和建模过程中没有考虑路由回路产生,路由出错处理机制的问题,重新设计算法流程,弥补算法设计的不足使其更适用于车载网的环境,并重新对该算法建模。随后,文中分析了AODV算法,指出采用跳数作为链路度量寻找最短路径的不足之处。考虑车载网的特点,利用模糊逻辑估计两个节点之间的链路质量,对Q-Learning进行改进。改变传统Q-Learning中折扣率不能根据具体的实际情况发生自适应变化的缺点。利用模糊逻辑估计链路质量并用得到的值替代Q-Learning中的折扣率。用Q-Learning算法对AODV算法进行改进,同时重新对该算法进行建模。文中分析了车载网中负载均衡的问题和AD-AODV负载均衡算法的不足之处。指出AD-AODV算法采用的链路度量方式的缺陷,针对该问题提出新的链路代价函数。同时对AODV算法进行改进,以实现负载均衡的策略并对改进后的算法重新建模。仿真实验结果表明所设计的算法与其它算法相比较具有较好的性能。
[Abstract]:In recent years, with the development of cloud computing, big data and other emerging technologies, the concepts of intelligent city and intelligent transportation have also come into people's view. The vehicular network attracts people's attention with its unique charm and broad market prospect. According to the characteristics of the vehicle network, this paper analyzes Q-Learning, ant colony algorithm and fuzzy logic reasoning, as well as some existing routing algorithms. This paper studies the application of Q-Learning and Ant Colony algorithm in vehicle network routing algorithm, analyzes Q-ABR algorithm, and points out the deficiency of pheromone updating method. Based on this, an improved method of pheromone updating is put forward. In the process of algorithm design and modeling, the routing loop generation and routing error handling mechanism are not considered in the process of algorithm design and modeling, and the algorithm flow is redesigned. The algorithm is more suitable for the vehicle network environment, and the algorithm is modeled again. Then, the AODV algorithm is analyzed, and the deficiency of using hop number as the link metric to find the shortest path is pointed out, and the characteristics of the vehicle network are considered. Using fuzzy logic to estimate the link quality between two nodes, Q-Learning is improved. The disadvantage of changing the discount rate in traditional Q-Learning can not be adapted to the actual situation. The link quality is estimated by fuzzy logic and the obtained value is used to replace the discount rate in Q-Learning. The AODV algorithm is improved, The problem of load balancing in vehicle network and the deficiency of AD-AODV load balancing algorithm are analyzed in this paper. The defect of link measurement method used in AD-AODV algorithm is pointed out. To solve this problem, a new link cost function is proposed, and the AODV algorithm is improved. The simulation results show that the proposed algorithm has better performance compared with other algorithms.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495;TN929.5

【参考文献】