文献类型：期刊文献

英文题名：The parallel optimization of network bandwidth allocation based on generalized particle model

作者：Shuai, Dianxun[1,2]; Feng, Xiang[1]

机构：[1]E China Univ Sci & Technol, Dept Comp Sci & Engn, Shanghai 200237, Peoples R China;[2]Tsinghua Univ, State Key Lab Intelligence Technol & Syst, Beijing 100080, Peoples R China

年份：2006

卷号：50

期号：9

起止页码：1219

外文期刊名：COMPUTER NETWORKS

收录：;EI(收录号：2006189855508);WOS:【SCI-EXPANDED(收录号:WOS:000237499000002)】；

语种：英文

外文关键词：bandwidth allocation; duality model; multi-objective optimization; generalized particle model; distributed parallel algorithm; computer networks

摘要：This paper presents a new duality approach to optimize the resource assignment and bandwidth allocation in networks. which is based oil the generalized particle model (GPM). By the GPM, the network's optimization problem is transformed into the kinematics and dynamics of numerous particles in two reciprocal dual force-fields. Like the Kelly's duality model and Low's duality model related to Lagrangian multipliers methods (LMM), the GPM-based duality model also contains two distributed algorithms for resource supplies and resource demands, respectively. and can be carried out in parallel. Both the LMM-based and GPM-based duality models may not only embody the market mechanism between the supplies and demands very well, but also may describe the influence of congestion degree of networks on the resource allocation. But the GPM-based duality model is different from other duality models in terms of the basic principle, mathematical formalization, and properties. Features of the GPM-based duality model and the corresponding duality algorithm include a powerful processing ability in a complex environment that involves the priority, personality, autonomy, and interaction of different entities in networks. Furthermore, the GPM-based duality model can realize the optimization of multiple objectives, including the aggregate utility, the personal utility, the minimal personal utility, the resource utilization, and the users' satisfactory degree. The proposed approach also has the advantages in terms of the higher parallelism, lower computation complexity, and the ease for hardware implementation. The properties of the GPM-based duality model, including the correctness, convergency and stability, are discussed in detail. The numerous simulations oil the network bandwidth allocation have shown the effectiveness and suitability of the proposed approach. (c) 2005 Elsevier B.V. All rights reserved.