In this thesis, a model for available bit rate (ABR) traffic class in asynchronous transfer mode (ATM) network congestion control with explicit rate (ER) feedback is introduced. To control the congestion, a optimal controller is proposed. First, the c...
In this thesis, a model for available bit rate (ABR) traffic class in asynchronous transfer mode (ATM) network congestion control with explicit rate (ER) feedback is introduced. To control the congestion, a optimal controller is proposed. First, the controller is applied to a global model, which uses linear matrix inequality (LMI) theory. Second, we apply the reduced-model to the controller, which uses weakly coupled and singular perturbation theory via the use of the bilinear transformation. The bilinear transformation can convert discrete Riccati equation into continuous Riccati equation. So, the calculation of Riccati equation be simplified. The proposed method has the rate of convergence of , where represents a small coupling parameter.
To apply weakly coupling and singular perturbation theory reduce the solving time of Riccati equation and simplified the calculation. So it makes the real-time transmission of ATM possibly.
To show the experimental feasibility and verification of the proposed controller, the controller is applied to a bottle-neck congested node simulation. The simulation shows that the proposed optimal controller guarantees the stability and Qos(Quality of service).