The paper shows how to encode Multi-Agent Path Finding problem as a sequential planning problem, specifically, how to encode various collision constraints and parallel actions.