Achieving High Performances for Nonuniform Traffic in an Input-Queued Switch with Chaotic Neural Network
Head-of-line blocking limits the throughput of an input-queued switch with first-in-first-out (FIFO) queues. Neural networks have been used to achieve high throughput, but when the traffic is nonuniform, these algorithms performance deteriorates and some queues may be starved in certain cases. Maximum weight bipartite matching algorithms, such as longest queue first (LQF) and oldest cell first (OCF) are introduced for nonuniform traffic. Compared with neural networks, the chaotic neural network has higher ability of searching for globally optimal or near-optimal solutions. In this paper, we combine the chaotic neural network with OCF, introducing waiting times of cells as a component of the energy function of the network. A method for adjusting the parameters of the network is proposed Computational simulation results show that our algorithm achieves higher performance for nonuniform traffic no matter it is admissible or inadmissible. Not only is the cell delay decreased greatly, but also the throughput is increased.
Chen Wenxia Zheng Junli
Department of Electronic Engineering Tsinghua University Beijing, China 100084
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
134-139
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）