|Abstract||The aim of this research is to model and simulate the recent and ongoing COVID-19 pandemic in terms of virus contagiousness among mixed groups of patients, carriers, and unaffected populations. The focus is on closed environments such as stores or schools that are typically ideal for the propagation of infectious pathogens. This research work utilizes real data from the State of Qatar to model and predict the behavior of COVID-19 as it spreads within human gatherings.
This work proposes the infection model SEIP (Susceptible-Exposed-Infected-Protected) developed to forecast the propagation of COVID-19 over time. The prediction model is applied to simulate different environments of human gatherings for the viral transmission under different biological factors. Applied machine-learning techniques based on reinforcement-learning algorithms trains smart agents who mimic the behavior of their human counterparts. Added 3D visualization, by harnessing the power of Unity 3D, further boosts the usability and appeal of the simulation.
The resultant simulation is customizable and extendable to simulate a myriad of possible pandemic scenarios and evaluate different potential safety control measures. Ultimately, we aim to equip authorities with a powerful tool to aid in decision making about future spreading of the virus under different controllable lockdown scenarios.