[=;190557]
Looking good. 
How big(in kilo/megabytes) is your pathfinding blueprint?
[/]
Thanks! I’m unsure. I’ll check when I have the opportunity. However, the blueprint is still undergoing big changes, and I know I’ll be deleting and adding a lot of paths.
How big is your blueprint? Do you have everything contained within a single blueprint, and what sort of blueprint is it? In my case almost everything is contained within an actor that has a single instance in the scene. All other blueprints also do most of their functions through this blueprint, using blueprint communications.
[=;190557]
Are you going to add any low-level AI stuff like deciding between 2 possible objectives when making a path? For instance, you tell the unit to move from one continent to another, so it decides which of 2 ports is closest or and/or best, and then chooses one to use for the water crossing. Then it would have a similar scenario when choosing a path to the best landing zone(s).
That sounds specific, but I think you could make it pretty generic with a multi-objective pathfinding process so that people could just assign objectives and then weight those objectives based on whatever factors.
[/]
That’s actually my next planned feature, and probably the biggest reason why I switched to using Dijkstra’s algorithm exclusively. All I have to do is add the location and attributes of relevant actors encountered during the path finding search. From these I can create an array of possible targets and then choose the one that fits certain criteria, such as the enemy with the lowest health, the closest city that has a port etc.
A quick note I’d like to add is that the blueprint no longer needs a physical grid of actors to function. Everything works just as well with just a single, solid plane, for instance. In the future I might add the possibility of doing pathfinding on uneven surfaces, perhaps with a grid projected onto a landscape mesh. That’s a bit down the line, though.