Swarm intelligence based robust active queue management design for congestion control in computer network

Dissertant

Khalid, Karam Samir

Thesis advisor

Ali, Hazim Ibrahim

University

University of Technology

Faculty

-

Department

Department of Control and Systems Engineering

University Country

Iraq

Degree

Master

Degree Date

2013

English Abstract

The congestion refers to a loss of network performance when a network is heavily loaded which leads to data loss and large delays.

In this work, three robust control algorithms are proposed to work as an Active Queue Management (AQM) which is an effective mechanism for congestion control problem.

These algorithms are; conventional H∞ controller, robust Particle Swarm Optimization based PID (PSOPID) controller and robust Ant Colony Optimization based PID (ACOPID) controller.

The design procedure of conventional H∞ method is complex and difficult to fine tuning the weighting functions in order to achieve the robust performance and robust stability.

Furthermore, the order of the resulting conventional H∞ controller is high, therefore; the Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) methods are used to tune the PID controller and the weighting functions parameters.

The parameters of the PID controller in the two types of controllers are tuned subject to H∞ constraints to achieve the required robustness of the network.

The robust PSOPID and ACOPID controllers can achieve desirable time response specifications with simple design procedure and low order controller in comparison to the conventional H∞ controller.

Low and wide ranges of system parameters change are used to show the robustness of the proposed controllers.

The ability of the proposed controllers to meet the specified performance is demonstrated using MATLAB 7.11.

On the other hand, to verify the effectiveness of the proposed controllers, a modification by adding the proposed control algorithms to the Ns2 (network simulator) function to perform the nonlinear simulation is done.

Finally, it is shown that the proposed robust ACOPID controller can achieve rise time = 0.408 sec., settling time =0.77 sec., mean =195.6 and standard deviation = 33.09, which are more desirable performance specifications compared to those obtained by the proposed robust PSOPID controller and the controllers that have been designed in previous works.

Main Subjects

Engineering & Technology Sciences (Multidisciplinary)

Topics

• Management
• Computer networks

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-418059