Manejo optimo y viable en modelos epidemiologicos de dengue

Abstract : Human epidemics are an important problem of public health in the world. Mathematical modelling is part of the instruments to fight them. The thesis «Optimization and viable control of epidemiological models of dengue» deals with the case of the dengue, an endemic disease in Colombia. The document is organized in two parts, a more theoretical Part I, and a Part II centered on applications. In Part I, the dynamics of propagation of an infectious vector-borne disease (such as dengue, transmitted by mosquitoes) is represented by systems of differential equations, connecting populations of individuals and vectors. We consider the epidemic model of Ross-Macdonald, and an endemic model of SIR-SI type. After the stability analysis of these dynamical models, this work brings forward two original theoretical contributions. The study of the transitory behavior deals with the control of an epidemic episode in its acute phase, before its possible asymptotic extinction. We look, by playing on the variable of control that is the mortality of the vector, to maintain human infected proportion (state) under a given threshold for all times (viability constraint). By definition, the viability kernel is the set of all initial states for which there is at least a trajectory of controls which allows to satisfy the viability constraint. Our main contribution is a complete description of the kernel. We discuss possible viable controls, whose application guarantees the satisfaction of the constraint. Then, we analyze two problems of optimal control. First problem is concerned with handling of an epidemic outbreak over a short timescale. Second one deals with an endemic infectious disease over a longer scale, where births and deaths within both populations (human and vector) are taken into account. We determine the necessary conditions of existence of an optimal solution by using the maximum principle of Pontryagin. We also tackle the case of resources limited over the time span. In part II, we apply the theoretical approaches of part I to the management of episodes of dengue in the city of Santiago of Cali. We estimate the parameters of the models by least squares, with the data supplied by the Program of epidemiological vigilance of the Municipal Secretariat of Public Health. We calculate numerically the viability kernel, fitted to the data observed during the epidemic episodes of 2010 and 2013.As for optimal control, we use the traditional forward-backward sweep algorithm, and compare several alternatives for the chemical control of the mosquito. The best strategy is a combination of spraying of an insecticide of low lethality, together with implementation of protective measures, which moderately reduce the biting rate of the mosquitos. Finally, we tackle the problem of dynamic control of the dengue under uncertainty. We develop a Ross-Macdonald model at discrete time with uncertainties. The robust viability kernel is the set of all initial states such as there is at least a strategy of insecticide spraying which guarantees that the number of infected people remains below a threshold, for all times, and whatever the uncertainties. Under proper assumptions on the set of scenarios of uncertainties (corresponding to temporal independence), an equation of dynamic programming allows to numerically calculate kernels. Having chosen three nested subsets of uncertainties a deterministic one (without uncertainty), a medium one and a large one we can measure the incidence of the uncertainties on the size of the kernel, in particular on its reduction with respect to the deterministic case (without uncertainty)
Document type :
Theses
Complete list of metadatas

Cited literature [43 references]  Display  Hide  Download

https://pastel.archives-ouvertes.fr/tel-01376000
Contributor : Abes Star <>
Submitted on : Tuesday, October 4, 2016 - 9:29:09 AM
Last modification on : Monday, October 28, 2019 - 2:28:11 PM
Long-term archiving on: Thursday, January 5, 2017 - 12:57:16 PM

File

TH2015PESC1034.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01376000, version 1

Collections

Citation

Lilian Sofia Sepulveda. Manejo optimo y viable en modelos epidemiologicos de dengue. General Mathematics [math.GM]. Université Paris-Est; Universidad nacional de Colombia, 2015. Español. ⟨NNT : 2015PESC1034⟩. ⟨tel-01376000⟩

Share

Metrics

Record views

266

Files downloads

238