This is a great write up. As a software (but not game) dev I’m always fascinated by the performance challenges faced by game devs, and the clever hacks they come up with. Presumably the flow field is only a worthwhile optimisation if the number of grid cells is fewer than the number of rats? Since I imagine it’s doing the same sort of path finding for each cell as it would otherwise be doing for each rat.
Yes I think it's worth saying that methods like this one are more effective at scale. While it's not necessarily a 1-to-1 mapping, if the number of active 'agents' moving around in the environment is not that large, then it may be more prudent to rely on a simple pathfinding algorithm (say A*) or use a navigation mesh tool that's built into the game engine. Given if we have a smaller number of moving agents, we're probably going to be placing a much stronger emphasis on ensuring their movements look as smooth as possible.
This is a great write up. As a software (but not game) dev I’m always fascinated by the performance challenges faced by game devs, and the clever hacks they come up with. Presumably the flow field is only a worthwhile optimisation if the number of grid cells is fewer than the number of rats? Since I imagine it’s doing the same sort of path finding for each cell as it would otherwise be doing for each rat.
Yes I think it's worth saying that methods like this one are more effective at scale. While it's not necessarily a 1-to-1 mapping, if the number of active 'agents' moving around in the environment is not that large, then it may be more prudent to rely on a simple pathfinding algorithm (say A*) or use a navigation mesh tool that's built into the game engine. Given if we have a smaller number of moving agents, we're probably going to be placing a much stronger emphasis on ensuring their movements look as smooth as possible.