基于2-阶段成本优化的多播容量供应

Multicast capacity provision based on 2-stage cost optimization

针对媒体内容提供商采用网络编码的多播最优化容量供应问题,提出一种基于2-阶段的成本优化方案。在第一阶段,网络容量购买基于对未来客户集的预测,这相当于内容提供商在与因特网服务提供商进行初始服务级协议谈判期间预定的容量;在第二阶段,一旦明确知道客户集,就追加购买服务所有客户所要求的额外容量;为指导第一阶段的容量购买决策,提出2种有效的近似算法,使两个阶段的总体成本在预期中最小化。仿真结果表明,该成本优化算法有效,且非常接近甚至优于最优解。

Aiming at the problems of the multicast optimal capacity provision when media content providers adopt network coding multicast,and a cost optimization scheme based on 2-stage was proposed.In the first stage,the network capacity purchase was based on the prediction of future customer sets,which was equivalent to the capacity reserved by the content provider during the initial service-level agreement negotiations with the Internet service provider.In the second stage,once the set of customers was known unambiguously,the recourse action at this stage was to purchase the additional capacity required to serve all customers.To guide the capacity purchase decisions in the first stage,two effective approximation algorithms were proposed to minimize the total cost of the two stages in expectation.The simulation results show that the proposed cost optimization algorithm is not only effective,but also very close to,even superior to the optimal solution.