This thesis examines the problem of regulating available bit rate (ABR) traffic in an ATM network. The ABR service category is defined in the ATM standards, so as to provide economical support for applications having minimal constraints for delay and throughput. A unique feature of ABR service is the requirement that sources be regulated using information from distant ATM nodes. This feature is problematic for source control schemes when the time-delays incurred in the feedback path and the temporal variation in the link capacity at the ATM node are to be considered. To utilize the link fully, accurate information about the network status has to be provided to ABR sources. In addition, sources should adjust their rates promptly according to this feedback information. Therefore the problem solution will require techniques that offer robust coordination between the sources and the nodes.
We examine the applicability of optimal control theory to this problem and present a design of a feedback control system for use in ABR service. We propose a source rate control algorithm which uses an ARMA (auto-regressive moving average) model for aggregate input rate into a buffer. The proposed source rate control algorithm allows for the coordination of the traffic between sources sharing a congested node. In this algorithm the weighted sum of past buffer occupancy and source rates are used to compute the new input rate.
A pole-placement technique is applied to choose the initial gains in the controller. These control gains are adjusted using a Linear Quadratic Tracker. As an aid to this end the system is reformulated in terms of state variables. Introducing a quadratic performance index, we formulate the problem as one of optimal linear quadratic tracker design. The coefficients of the controller are chosen subject to the minimization of a quadratic performance index. The minimization of the weighted sum of the control input and error energy is performed. The algebraic matrix Riccati equation is utilized to find the optimal control law of the closed-loop control system considered.
We demonstrate the efficiency of the algorithm based on the
From our numerical results, it is shown that the parameters of the
model can be obtained by minimizing the weighted sum of the input
error energy subject to the system equation constraint.
By adjusting the weight of the error energy the rise time can be
The minimum error energy resulted in the best tracking performance.