Development of model predictive controller for congestion control problem

Abstract EN

Nowadays congestion in computer networks is pointed out as an important and a challenging problem.

TCP (Transmission Control Protocol) has the mechanism to avoid congestion in computer networks.

TCP detects congestion by checking acknowledgements or time-out processing and adjusts TCP window sizes of senders.

However, this control method shows low efficiency in communications because it is based on a mechanism that avoid congestion after congestion once appears in computer networks.

TCP random early detection RED is another popular congestion control scheme.

The fundamental idea behind this control algorithm randomly drops the incoming packets proportional to the average queuing length and to keep the queuing length to a minimum.

To achieve high efficiency and high reliability of communications in computer networks, many control strategies based on advanced control theories have been introduced to tackle the congestion problem.

Model Predictive Control (MPC) is the only practical control method that takes account of system constraints explicitly, and the only ‘advanced control’ method to have been adopted widely in industry.

MPC is a model-based method which uses online optimization in real time to determine control signals.

The solution to optimization problem is usually formulated with the help of a process model and measurements.

At each control interval, an optimization algorithm attempts to determine the plant dynamics by computing a sequence of control input values satisfying the control specifications.

In this work, a planning strategy based on MPC will be developed for congestion control problem.

A "preset controllers" approach will be introduced for such application.

The effectiveness of considered controller will assessed in terms of how well it could show good tracking performance, maximizing the utilization of the available bandwidth and to what extent it could cope with system uncertainties.