| 摘要: |
| Ad Hoc网络存在着无线多跳连接、节点移动这两个本质的特点.前者引起了与固定网络截然不同的信息流竞争新特点,后者导致了网络状态不断发生变化.首先,在采用链路干扰集描述信息流竞争特点的基础上,将小时间段内网络状态不变的拥塞控制问题表达成非线性优化问题;其次,运用基于对偶分解理论的价格协作法PCA(price cooperation approach)求解该优化问题,构建了一个基于链路干扰集的价格框架.同时,运用队列长度监控、邻居集合近似和HELLO捎带信息这3种技术将PCA转化成在实际Ad Hoc网络环境 |
| 关键词: Ad Hoc网络 拥塞控制 对偶分解 滚动优化 价格协作 |
| DOI： |
| 分类号: |
| 基金项目:Supported by the National Natural Science Foundation of China under Grant Nos.60702081, 60604029, 60473038 (国家自然科学基金); the Joint Funds of NSFC-Guangdong under Great No.U0735003 (国家自然科学基金委-广东省联合重点基金项目); the Natural Science Foundation of Zhejiang Province of China under Grant Nos.Y107309, Y106384 (浙江省自然科学基金) |
|
| Congestion Control for High Dynamic Ad Hoc Networks: Price Cooperation and Receding Optimization |
|
XU Wei-Qiang,WU Tie-Jun,WANG Ya-Ming,ZHANG Yun-Hua,CHEN Ji-Ming
|
| Abstract: |
| In Ad Hoc networks, there exist two fundamental characteristics: Multi-Hop wireless-based transmission and node mobility. The former results in different contention relationship among flows from that in wireline networks, and the latter leads to the time-varying network situations. Firstly, on the basis of the link's interference set depicting the characteristics of the contention relationship, the congestion control problem for small time interval is formulated as a nonlinear optimization problem. Secondly, by using the dual decomposition theory, a price cooperation approach (PCA) is proposed to solve this optimization problem. In PCA, a price framework based on the link's interference set is built. Meanwhile, to implement PCA in realistic ad hoc environment, three deployment techniques are proposed: Queue Size Monitoring, Neighbor Set Approximation, and HELLO-based Message Piggybacking. Otherwise, the network status detection and receding optimization is introduced to deal with the uncertain changes of network situations, and an adaptive optimization strategy (AOS) is proposed correspondingly. The simulation results in MATLAB environment show that AOS has better performance of adaptation to time-varying network situations than PCA. The simulation result in NS2 environment show that PCA and PCA+AOS significantly outperform TCP, ATCP, and ATP in many important performances, including throughput, packet drop ratio, and fairness, under a variety of scenarios and mobility models. |
| Key words: ad hoc network congestion control dual decomposition receding optimization price cooperation |
