Our life is influenced by COVID-19 for more than half a year now. What is known about the virus is that it spreads in a very non-uniform way. Some infected people do not transmit the disease not even to their family, while others may infect hundreds of others. This feature makes this disease a difficult target for calculation and simulation. Recently we have developed a model based on the SIR model which includes many new features the virus has and policies the governments employed to contain the infection. The code runs efficiently on NVidia graphic cards. The task of the thesis will be to further improve the model to include more realistic social effects and compare the results to the publicly available data.