Adaptive Pathfinding

Hi,

So imagine I have a complex of rooms. Each room is joined by a door that can be locked or unlocked. Door states can change at any time.

Now I have a bunch of agents that I want to path-find across these rooms. The agents know the layout of the complex, they have the ‘blueprint’, so should plan the most optimal route to the target.

The agents initially assume all the doors are open and plan accordingly. However as they come across a locked door they need to re-plan their route, remembering the state of the locked door (however it may be unlocked/locked on them again in the future).

So whats the best way to achieve this? I was thinking of remembering each door node state for each agent and then setting the weight according to that agents ‘memory map’ before I call find route. The problem with that approach is that I can only have one path running at a time, as the weights will be set on a per-agent basis.

There will be many agents (50+) so having a separate path-finding map per agent would be prohibitive.