Do they factor in natural evolutionary pressure and change of the virus?

Not at all. Meta population models use some type of stochastic discrete PDEs and agent based models just infection probabilities. This is an over simplification but just so you have an idea. You can get the gritty details in the papers above.