Congestion Prediction Control Of ATM Network Based On Particle Filter

Due to the abruptness and time-variable of the network traffic in Asynchronous Transfer Mode (ATM) networks, the source can not respond as quickly as the network status changes, which will result in massive cell loss. The congestion of information is the main reason that affects the quality of service (QoS). Research on congestion control has important theoretical and practical value. Therefore, the research of the thesis is from the perspective of prediction control, based on the analysis of current congestion control algorithms.Considered the statistical properties of the ATM network sources data, with the excellent prediction performance of particle filter algorithm, the thesis researched a network congestion prediction mechanism based on particle filter. The main works are as follows:(1) Based on the analysis of the causes, characteristics, judgment means, and design objectives of network congestion, the prediction decision-making mechanism of network congestion is proposed.(2) The queue model for ATM network system was analyzed. The cell loss rate was specified as the measurement of network congestion control target. Particle filter was used as the prediction process, and combined with network congestion control process, do a good job in reducing the cell loss rate of the source.(3) Considered the balance of the diversity and effectiveness of prediction particles, the genetic mechanism is applied to its re-sampling module, to improve the performance of prediction control further.The core idea of this thesis is to apply particle filter and genetic re-sampling particle filter algorithm to ATM network congestion prediction control, in order to get a lower cell loss rate.The simulation result shows that, Compared with another kind of prediction scheme with fuzzy neural network method, this scheme is able to adjust source rate at real-time and self-adapt, and reduce cell loss rate greatly. The genetic re-sampling particle filter has a lower cell loss rate, gets better congestion control results, and ensures the quality of network services.