I will try to make clear analogy:
There is a city with N destinations. It is represented by weighted and directed graph where weights are distances as minutes.
There are 2 people which don't want to be in same destination at same time. They are located in different destinations. They are going to go to another destinations using shortest paths without being in same destination at all. They will stay M minutes in each visited destination.
How to find shortest path while keeping them distant?
NOTE: I have looked shortest path algorithm, traveling salesman algorithm and thier variations. But I can't figure out how to solve it effectively. Finding all paths and comparing intersections doesn't look cost efficient.