网络资源并行优化的纵横复合弹簧网络方法

A Crossbar Composite Spring-Net Approach to Parellel Optimization for Computer Networks Resources

具有复杂社会动力学特性的计算机网络中 ,存在着许多难以求解的问题 .该文提出一种新的纵横复合弹簧网络模型和方法 ,对计算机网络资源的分布式优化问题建立了相应的数学物理模型及分布平行算法 ,将网络资源优化问题求解过程转变为对应的复合弹簧网络形变过程 .该模型和方法能刻画网络实体之间并发的复杂的社会交互行为 (竞争、协作和诱导等 ) ,能描述网络实体随局势的变化各自采取的动态策略和自治行为 ,并能体现出市场价格机制的作用 .

This paper presents a new parallel approach based on crossbar composite spring net model, by which solving the resource assignment in computer networks is transformed into parallelly calculating the deformation of the corresponding crossbar composite spring net. The approach can deal with various complicated social interactions occurred in computer networks and describe the behavior dynamics involved in autonomous strategy adjustment among computer network entities, which may be formidably difficult problems to solve by using other traditional methods. The important role of market pricing mechanism during the optimization of resources assignment for computer networks can well be embodied by the proposed approach. An application of this approach to bandwidth assignment in broad networks is shown.