SEIRD COVID-19 Formal Characterization and Model Comparison Validation

Abstract

Based on a SEIRD model (Susceptible, Exposed, Infective, Recovered and Deceased) for COVID-19 infection with a new parametrization using a high infection rate, and a low fatality, we define the model in System Dynamics, Python, and Specification and Description Language (SDL). The different implementations obtained can be improved depending on the capabilities of the approach and, more interestingly, can be used to improve the Validation and Verification processes. In this paper, we are focused on describing how this comparison with other models’ validation processes allows us to find the parameters of the system dynamics model, hence the parameters of the pandemic. This is a crucial element, specifically in this case, because the data are not complete or validated for different reasons. We use using existing data from Korea and Spain and showing that the proposed method and the obtained parametrization for the model fit with the empirical evidence. We discuss some implications of the validation process and the model parametrization. We use this approach to implement a Decision support system that shows the current pandemic situation in Catalonia.