多目标粒子群优化的复合型组播拥塞控制  

Composite Multicast Congestion Control by Multi-objective Particle Swarm Optimization

作　　者：麻海圆[1,2] 孟相如[1] 马志强[1] 马森[2] 

机构地区：[1]空军工程大学电讯工程学院 [2]中国人民解放军94303部队

出　　处：《应用科学学报》2011年第5期459-466,共8页Journal of Applied Sciences

基　　金：陕西省自然科学基础研究计划项目基金(No.SJ08F14;No.2009JQ8008)资助

摘　　要：为提高异构环境下大规模动态组播的吞吐量并减小速率抖动,提出一种复合型组播拥塞控制机制(composite multicast congestion control,CMCC).该机制在基层和增强层使用不同的速率调节方案:基层使用单速率机制控制发送速率,防止接收能力较弱的接收端被"排挤"出组;增强层使用最优化的方法分配速率以提高组播吞吐量并平滑发送速率.CMCC以最大化全局满意度和最小化速率抖动为目标将增强层速率的分配归纳为多目标优化问题,并选用收敛速度快、参数少且计算量小的多目标粒子群优化算法搜索最优解.仿真结果表明,CMCC能有效提高组播吞吐量,平滑发送速率,且计算量小,实时性高,适用于具有异质接收端的大规模组播.To enhance throughput and smoothen the sending rate of large scale multicast in a heterogeneous environment, a composite multicast congestion control （CMCC） scheme is proposed. The rate in basic layer is controlled by the lowest receiver in a single-rate way, while the rates in enhanced layers are allocated by an optimizer. Aimed at maximizing global satisfaction and minimizing jitter of sending rates, rate allocation of layers is formulated as a multi-objective optimization problem. The multi-objective particle swarm optimization （PSO） algorithm is used to solve the multi-objective problem. The algorithm has advantages of fast convergence, small number of parameters and low computation complexity. Simulation results show that CMCC can enhance throughput as compared with traditional hybrid multicast congestion control scheme. It also has a smooth sending rate in streaming applications.

关 键 词：组播 分层组播 拥塞控制 多目标优化 粒子群优化 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]