Pinning control of disease networks

The modelling of contagion spread on contact networks provide valuable insights to epidemiologists and policymakers trying to control and eradicate diseases. This thesis proposes, implements and analyses a methodology for inserting disease contact networks of HIV into feedback control loops and applying open-loop pinning control to their nodes. Pinning control aims to medicate only a portion of an entire network in order to achieve the same outcomes that would be seen when all nodes are controlled. The control loops are simulated using networks ranging from size N = 100 nodes to N = 10000 nodes. Simulations aim to control the average maximum incidence in the networks by first estimating the reference average transmissibility from the statistical physics technique known as bond percolation. Once the average transmissibility is known, node-, network- and population mass-action models can be measured for incidence. Two selective pinning control strategies, namely proportional feedback and nonlinear model predictive control (NMPC), are compared with one another and also with a random pinning strategy. The budget, measured in quality-adjusted life years (QALYs), is added to the cost-function for NMPC control. It is shown that budget can indeed be controlled while incidence varies, while incidence may be controlled as budget varies. Pinning control of disease networks is a feasible methodology to analyse the future and steady-state outcomes of interventions in fast-spreading (high-risk) disease contact networks.

Grantor dc:publisher

University of Pretoria

Year dc:date.issued

2015

Advisor dc:contributor.advisor

• Craig, Ian K.

• UCTD

Statement dc:rights

• © 2016 University of Pretoria. All rights reserved. The copyright in this work vests in the University of Pretoria. No part of this work may be reproduced or transmitted in any form or by any means, without the prior written permission of the University of Pretoria.

Language dc:language.iso

en

Dc Identifier Other

A2016

OAI identifier oai:identifier

oai:repository.up.ac.za:2263/56133