I recently got triggered to recreate the A* pathfinding algorithm. To be fair - you can use the findshortestpath node but I love to digg into these algorithm and challenge myself to find a solution.
But there is a bigger idea behind this. I wanted to do something similar for a while now. It is one puzzle piece to create crowd effects I was facinated with.
For example this scene from the upcoming game cyberpunk 2077
Natural behaviour for these kind of crowds are difficult and not too much information out there.
This was a very good and interesting presentation by RISE on a Hive event.
They aswell used paths for the individual agents aswell as the fuzzy logic. The combination delivered good results. Of course the assets they created are company secrets.
But I am now a few steps closer.
The path constructed on a grid layer passes the information down to create a curve path on the "level".
And finally after a bit of reconfiguration I used the fuzzy example file to apply my created path. Now it is time to capsule it as an asset which can be used to create paths foreach agent.
The basic algorithm will find a place in the tool asset while the more detailled approach will be the foundation of a dop asset.