机会网络中高效路由协议的研究与实现

发布时间：2018-08-15 18:58

【摘要】：机会网络的概念来自容迟网络,是一种新型的移动自组织网络,网络中的节点利用彼此运动所创造的接触机会进行消息的传递,通信双方无需事先创建完整的端到端的通信链路。网络内部节点的频繁移动,致使临时建立起的通信连接经常性中断。因此,针对机会网络的特性,设计适用于该网络的高效的路由算法显得尤为重要。本文针对机会网络广泛使用的路由算法进行了详细的探讨和研究工作,设计并实现了新的路由算法。以高效性作为路由算法设计的出发点,本文提出一种利用Prophet投递预测函数和马尔可夫模型改进Spray and Wait的PMSF(Prophet and Markov model-based Spray and Forward)算法,该算法在Spray阶段充分考虑中继节点的传递潜能及传输可靠性,将消息副本根据Prophet算法中的投递概率进行相应的分配,使消息的传输决策更为明智,在传输预测值的更新过程中引入路由可靠性的设计,使得消息能够在最稳定且高效的链路上进行传输。PMSF算法将Wait阶段改为主动路由的多跳转发阶段,并利用马尔可夫模型对节点相遇时间间隔做出合理的预测,继而把能在最短时间内与目的节点相遇的节点作为下一跳转发节点。在整个路由过程中,引入基于ACK确认的缓存管理策略,在一定程度上有效缓解了网络拥塞的问题。通过大量仿真实验来验证算法的有效性和效率,结果表明,该路由算法在消息投递率,投递延迟和路由开销方面能够表现出更好的性能。
[Abstract]:The concept of opportunistic network comes from delay-tolerant network, which is a new type of mobile ad hoc network, in which nodes make use of the contact opportunities created by each other's movements to deliver messages. There is no need for both parties to create a complete end-to-end communication link in advance. Frequent movement of nodes within the network results in frequent interruptions of temporarily established communications connections. Therefore, it is very important to design an efficient routing algorithm for opportunistic networks. In this paper, the widely used routing algorithms in opportunistic networks are discussed and studied in detail, and a new routing algorithm is designed and implemented. Taking high efficiency as the starting point of routing algorithm design, this paper proposes a new PMSF (Prophet and Markov model-based Spray and Forward) algorithm which uses Prophet delivery prediction function and Markov model to improve Spray and Wait. In the Spray stage, the transmission potential and transmission reliability of the relay node are fully considered, and the message replica is allocated according to the delivery probability in the Prophet algorithm, which makes the message transmission decision more intelligent. The design of routing reliability is introduced into the update process of transmission prediction value, which enables the message to be transmitted on the most stable and efficient link. The algorithm can change the Wait phase into the multi-jump phase of active routing. The Markov model is used to predict the time interval of the node meeting reasonably, and then the node that can meet the destination node in the shortest time is used as the next hop-forward node. In the whole routing process, the buffer management strategy based on ACK confirmation is introduced, which effectively alleviates the problem of network congestion to a certain extent. The effectiveness and efficiency of the algorithm are verified by a large number of simulation experiments. The results show that the routing algorithm has better performance in message delivery rate, delivery delay and routing overhead.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN929.5

【参考文献】