> It looks like you could model this naively with a discrete time Markov chain using a 3x3 stochastic matrix and three states: healthy, infected and deceased.

> It looks like you could model this naively with a discrete time Markov chain using a 3x3 stochastic matrix and three states: healthy, infected and deceased. Diffusion is a Markov process, yes. But you'd need three states per cell (not sure if that's what you meant by three states).