Applying perturbation analysis to traffic shaping

Authors:
D. C. Lee
Affiliations:
University of Southern California, Department of Electrical Engineering - Systems, 3740 McClintock Avenue, Los Angeles, CA 90089-2565, USA
Venue:
Computer Communications
Year:
2001

Citing 2
Cited 1

Optimization of queues using an infinitesimal perturbation analysis-based stochastic algorithm with general update times

SIAM Journal on Control and Optimization
Using perturbation analysis for gradient estimation, averaging and updating in a stochastic approximation algorithm

WSC '88 Proceedings of the 20th conference on Winter simulation

Parameter optimization of tandem queue systems with finite intermediate buffers via fuzzy simulation

Performance Evaluation

Quantified Score

Hi-index	0.24

Visualization

Abstract

This paper presents an idea of designing semi-open-loop traffic-shaping mechanisms for real-time network applications on the conceptual bases of perturbation analysis and stochastic optimization. A simple, representative traffic-shaping scheme is introduced, which ensures that the time between successive PDU entries into the network is not smaller than parameter @q. A stochastic algorithm based on an infinitesimal perturbation analysis is proposed to evaluate the optimal value of @q for a cost function, which takes into account the fact that the queueing delay in the network is costlier than the waiting time in the source premise. The waiting time in the source premise and the queueing delay in the network are modeled by the delays in tandem queues. This paper also illustrates the use of the renewal theory in stochastic optimization. It is suggested that the stochastic algorithm be used to track the optimal values of @q for time-varying bandwidth provision.