摘要
针对贪婪地理路由协议的贪婪转发策略在业务量比较大时系统吞吐量急剧下降、网络节点能量消耗严重不平衡,缩短了系统的使用寿命,提出一种基于部分状态可观测的马尔可夫决策理论的负载均衡贪婪转发策略,设定队列长度剔除门限值,有效地控制传输路径上节点的队列长度,通过避开负载较重的节点来实现网络负载均衡。仿真结果表明,该算法能降低端到端时延、减少数据分组的丢失、提高网络数据传输的实时性和可靠性。
Greedy geographic routing protocol greedy forwarding strategy, the system throughput decreased sharply when the business volume is large, the energy consumption of the network nodes seriously unbalanced, shorten the service life of the system. This paper presented a load balancing greedy forwarding strategy based on part state 6bservable Markov decision theo- ry, this mode set the queue length culling threshold value,it could effectively control node's queue length at the transmission path, achieve the network load balancing by avoiding a heavily loaded node. The simulation results show that the algorithm re- duces the end-to-end delay and data packets reduce loss, improve data transmission reliability and real-time on the network.
出处
《计算机应用研究》
CSCD
北大核心
2014年第5期1532-1535,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(61202490)
江西理工大学科研基金资助项目(jxxj12060)
江西省2013艺术科学规划项目(YG2013037)