Viable Control of an Epidemiological Model

Michel de Lara (1) , Lilian Sofia Sepulveda Salcedo (2)

1 CERMICS - Centre d'Enseignement et de Recherche en Mathématiques et Calcul Scientifique
2 UAO - Universidad Autónoma de Occidente

In mathematical epidemiology, epidemic control often aims at driving the number of infected individuals to zero, asymptotically. However , during the transitory phase, the number of infected can peak at high values. In this paper, we consider mosquito vector control in the Ross-Macdonald epidemiological model, with the goal of capping the proportion of infected by dengue at the peak. We formulate this problem as one of control of a dynamical system under state constraint. We allow for time-dependent fumigation rates to reduce the population of mosquito vector, in order to maintain the proportion of infected individuals by dengue below a threshold for all times. The so-called viability kernel is the set of initial states (mosquitoes and infected individuals) for which such a fumigation control trajectory exists. Depending on whether the cap on the proportion of infected is low, high or medium, we provide different expressions of the viability kernel. We also characterize so-called viable policies that produce, at each time, a fumigation rate as a function of current proportions of infected humans and mosquitoes, such that the proportion of infected humans remains below a threshold for all times. We provide a numerical application in the case of control of a dengue outbreak in 2013 in Cali, Colombia.

Mots clés: epidemiology; control theory; viability theory; Ross-Macdonald model; dengue

Domaines: Optimisation et contrôle [math.OC]; Systèmes dynamiques [math.DS]; Santé publique et épidémiologie

Michel de Lara, Lilian Sofia Sepulveda Salcedo. Viable Control of an Epidemiological Model. Mathematical Biosciences, 2016, 280, pp.24-37. ⟨10.1016/j.mbs.2016.07.010⟩. ⟨hal-01208484v2⟩