Numerical Exploration of the Parametric Identification and Delay Reconstruction Problem on Epidemiological Models Using Evolutionary Algorithms

The emergence of the COVID-19 pandemic in December 2019, caused by the SARSCoV-2 virus, has prompted extensive analysis using mathematical models like SIR (Susceptible, Infected, Recovered) and SEIR (Susceptible, Exposed, Infected, Recovered) to comprehend the dynamics of the disease. Recent advancements in model refinement include the incorporation of time delays to account for incubation and recovery periods, as well as adaptations to encompass the transmission by asymptomatic individuals. However, the precise identification of parameters in these models remains a challenging task due to incomplete and non-uniform data. This study delves into the utilization of numerical algorithms, specifically Particle Swarm Optimization and Genetic Algorithms, to optimize parameters and delay times in SIR models based on recent pandemic data. The performance of both algorithms is explored through the testing of parameter identification under varying levels of uncertainty in measures of infected patients. The results of this research aim to improve the understanding of the COVID-19 pandemic and contribute to the refinement of mathematical models for analyzing infectious disease dynamics.