基于梯度的DTN路由算法

A Gradient Based Routing Algorithm for Delay Tolerant Networks

针对容迟网络(DTN)中多副本消息传递造成网络资源浪费的问题,提出了一种基于概率和新鲜度的梯度路由(PFG)算法.该算法通过概率梯度和新鲜度梯度来构建目标节点的历史轨迹与目标节点信号覆盖所形成的势场.在节点相遇时,将消息传递给更新鲜、概率更高的节点,尽力避免网络资源的浪费.仿真实验结果表明,PFG路由算法在满足相同消息传递完成率和消息平均时延的条件下,可减少消息传递的副本数,提高系统性能.

In order to reduce the waste of network resources caused by multiple copies routing in delay tolerant networks （DTN） , a routing algorithm based on probability gradient and fresh gradient （PFG） is proposed. In the algorithm, the historical track of the destination node and the coverage potential field of the destination node are established with probability gradient and fresh gradient. When the two nodes meet, the message will be sent to the node which is much fresher and with higher probability to the destination node, it means that more waste of network resources can be avoided. Simulations show that the PFG can efficiently reduce data copies in network under the same message delivery rate and message average delivery delay, and hence improve the system performance.