基于福利经济学的网络资源分配模型

采用福利经济学的分析方法,提出网络资源分配的生产-消费模型,通过引入效用可能性边界和社会福利函数,得到资源分配问题的优化目标,指出当社会福利函数取得最大值时,网络系统性能取得效率和公平意义上的最优.通过对资源集和资源约束、用户集和服务需求进行数学抽象,将资源分配的概念模型转化为一个多约束优化模型NRAP.相对于Kelly无差别速率分配模型,本模型可以较完整地刻画网络传输服务的生产和消费特性,且能够反映用户在服务质量上的不同要求.分析说明了NRAP资源分配模型在实际网络中的应用及算法的实现机制.

Welfare-economic model of the network resource allocation

Following the methods of welfare economics, a production-consumption model for network resource allocation is presented firstly. The concepts of utility possibility frontier and social welfare function (SWF) are discussed. It is then illustrated that the network resource allocation problem gets its optimum solution in terms of efficiency and fairness when the corresponding SWF reaches its maximum value. Thereafter the conceptual model is transformed to a mathematical optimization model named as network resource allocation programming (NRAP), subject to multiple constraints including resources set, users set and service requirements. In comparison with Kelly's rate control model, NRAP lends itself to characterize particularly the production and consumption of network services as well as the user demands on quality of service (QoS). Finally, some rate (control) algorithms by which NRAP can be taken into practical networks are examined to show the (model's) rationality and feasibility.