Comment by theshrike79
Comment by theshrike79 3 days ago
Isn't the A* algorithm something that kinda helps with this? Basically you recalculate the route every "step".
Comment by theshrike79 3 days ago
Isn't the A* algorithm something that kinda helps with this? Basically you recalculate the route every "step".
I don't think so. A* just let's you add domain knowledge to the process of selecting the next candidate to explore.
But all (classic) pathfinding algorithms "teleport around" in the graph in some way, either switching between different candidate routes or backtracking from a dead-end. That makes it impossible to run the algorithm "live", I.e. have the agent navigate to a location as soon as the algorithm got to it.
A* still produces a full path ahead of time, it's just potentially more efficient to compute. Recalculating each step would be a layer on top of that.