【摘要】 提出了一种基于时间序列预测的延迟容忍网络路由算法.采用改进的加权移动平均时间序列预测算法,利用不断积累的历史链接信息,预测链路未来链接状态,构建路由表信息.该算法不要求节点具备先验知识,能随着网络拓扑的变化调整消息的传输路径,克服了同类算法需要大量网络链接先验知识的弊端.仿真结果表明,与最小期望延迟、最小估计期望延迟算法相比,该算法具有更高的传输成功率和更低的平均传输延迟.更多还原

【Abstract】 A routing algorithm based on time series prediction for delay tolerant network is proposed.The weight moving average method as one of time series prediction methods is improved,and the history connected information of nodes is used to predict the connection state in the future.The characteristic of the proposed algorithm include that it does not need any network connection information,and the routing path can change followed with the change of network topology.The proposed algorithm can avoid the shortcoming of existing algorithm which needs prior knowledge.Simulation shows that the proposed algorithm compared with minimum expected delay and minimum estimated expected delay have higher delivery ratio and lower average time.更多还原