Abstract
This study presents a novel parameter calibration technique applied to stochastic network models, specifically the Erdös-Rényi and Barabási-Albert models, to simulate disease transmission dynamics. Using the Differential Evolution algorithm for parameter calibration, we reasonably matched the initial infection peak of COVID-19 observed in the United States. In particular, the Erdös-Rényi network showed superior adaptability in capturing the infection curve. The research highlights challenges such as the variability of results due to the stochastic nature of the simulations and the complexity of parameter fitting. Advances in this area are expected to improve the robustness and applicability of stochastic network models.