Particle colony entity

TeamSangai

Hello, just wondering how feasible it would be to make a sort of 'living' super entity that subsists of groups of particles.
looking to try to do something like that for 'Ant' colonies in a Rain World style game.

decryptedchaos

I would say this will depend on the scale. the engine will gradually get slower the more entities that persist if you want them to in fact be true entitles and not simulated / fake ones.

In the past, Physics is what hindered entity performance the most, 4.4 (soon to be released) will inave Jolt Physics and that will bring with it a lot of performance uplift when dealing with entities

TeamSangai

decryptedchaos I would say this will depend on the scale. the engine will gradually get slower the more entities that persist if you want them to in fact be true entitles and not simulated / fake ones.

the ants themselves will mostly just be particles, moving around and only having gravity if they move out of a cluster's zone. along with existing for only moments at a time.
the ant's behavior and origin would be from said clusters, usually just about one main cluster, though can be more if it needs to split up (for example if a larger creature tramples on the colony and the ants need to attack that creature, then a new cluster could be created and 'stick on' that creature).

Edit: at least that's what my plan is.

decryptedchaos

TeamSangai the ants themselves will mostly just be particles, moving around and only having gravity if they move out of a cluster's zone. along with existing for only moments at a time.

Well, what i'm thinking is, AFAIK with physics, its either on or off.. you can create your own state machine in code that will flip the physics on or off depending on if it needs to interact, but doing this could have undesirable effects.

I'll let someone else chime if possible, but having a system where physics only applies for breif moments may have overhead issues, it could also cause odd behavior in runtime. you might be better off simulating gravity without physics, i'm not sure.

Zeftax

If it is too computationally intensive, maybe you could also try having sub-clusters (or maybe even the clusters directly, idk how big you want 'em) shaded to look like a bunch of ants as the physically-simulated particles, instead of simulating each ant individually. All the ants in the cluster will probably move together after all.