In this paper, we reflect upon control intervention practices habitually exerted by healthcare authorities in tropical areas that suffer from incidental outbreaks of dengue fever, in particular, the city of Cali, Colombia. Such control interventions, principally based on the insecticide spraying, are carried out sporadically in order to overcome an ongoing epidemic or at least to reduce its size. It is worth pointing out that control actions of this type do not usually account for sufficient budget because epidemic outbreaks are difficult to predict. In practical terms, these occasional control interventions are performed by spraying, as quickly as possible, all existing stock of insecticide (regardless of its lethality) and employing all available manpower. The goal of this paper is to design better strategies for insecticidebased control actions, which are capable of preventing more human infections at no additional cost, and to reveal the obsolescence of current vector eradication practices. Our approach relies on dynamic optimization, where the number of averted human infections is maximized under budget constraint and subject to a simple dengue transmission model amended with one control variable that stands for the insecticide spraying. As a result, we obtain structurally robust control intervention policies that demonstrate better performance and higher resilience to possible budget limitations than traditional modus operandi. K E Y W O R D Sdengue outbreaks, insecticide-based vector control, isoperimetric constraint, optimal control, optimization, Ross-Macdonald model Stud Appl Math. 2020;144:185-212.wileyonlinelibrary.com/journal/sapm

Dengue is an infection transmitted by mosquitoes and is present in all tropical and subtropical regions of the planet. In recent years, the transmission of the disease has increased, predominantly in urban and semi-urban areas, and has become an important public health problem. The National Health Institute (Instituto Nacional de Salud-INS) determined Cali to be the municipality with the most cases of dengue in Colombia. According to the INS, up to epidemiological week 18 (April 28 to May 4, 2013) 5 134 cases of dengue -and 171 cases of severe dengue- have been reported. This study presents a description of the Ross-Macdonald model, and qualitative analysis of this model, and an analysis of the sensitivity of the model to changes in its parameters. Based on the adjustment of the model obtained for cases that occurred in 2010, simulations of possible scenarios of epidemic outbreaks in the city of Cali are created and analyzed.

The classical Kermack-Mckendric SIR model for infectious disease transmission was used for studying the dynamics of HIV-AIDS in the city of Cali; individuals were classified as being susceptible, infected or recovered (SIR) on the assumption that sexual transmission would be the only means of transmission and that individuals would not recover or die. The model's parameters and basic reproductive numbers were estimated using information supplied by the Santiago de Cali Municipal Secretariat of Health, the Colombian Statistics Bureau (DANE) and the Municipal Planning department. Some scenarios were simulated to establish long-term disease trends. The system's equilibrium points were estimated and stability conditions analyzed finding that the current system had two equilibrium points: unstable, disease-free (E1) and stable, endemically asymptotic (E2). Taking information from 2008 as initial conditions, it was observed that the disease would tend towards equilibrium after a 100 year endemic. Simulations suggested that the disease would tend towards endemic equilibrium more slowly by reducing the probability of contact between susceptible and infected individuals and that the maximum number of infected and recovered could also become reduced.