This study utilizes modeling and simulation to analyze coronavirus (COVID-19) infection trends depending on government policies. Two modeling requirements are considered for infection simulation: (1) the implementation of social distancing policies and (2) the representation of population movements. To this end, we propose an extended infection model to combine analytical models with discrete event-based simulation models in a hybrid form. Simulation parameters for social distancing policies are identified and embedded in the analytical models. Administrative districts are modeled as a fundamental simulation agent, which facilitates representing the population movements between the cities. The proposed infection model utilizes real-world data regarding suspected, infected, recovered, and deceased people in South Korea. As an application, we simulate the COVID-19 epidemic in South Korea. We use real-world data for 160 days, containing meaningful days that begin the distancing policy and adjust the distancing policy to the next stage. We expect that the proposed work plays a principal role in analyzing how social distancing effectively affects virus prevention and provides a simulation environment for the biochemical field.
Keywords: COVID-19 epidemic; SIRD model; data-based learning; discrete-event model; simulation.