swarm behaviour

[EXit_W0und]

In case anyone is thinking of editing the waypointing or experimenting more with squad behaviour within the Spring engine: this might be a useful article.

http://www.environmentalgraffiti.com/sc ... swarms/748

[I2:Isaacment_Day]

[TheBigPK]

lol

but I can├óÔé¼Ôäót think of a positive situation where I would see a swarm of tiny robots and think to myself ├óÔé¼┼ôThank God the robot swarm is here.├óÔé¼┬Ø

I can...

[Dragon45]

Google "boids". This is all old shit.

[SwiftSpear]

There are acctually peer reviewed papers on AI pathing in a swarm enviroment... Suffice to say they would likely be more useful than that blog article

[KDR_11k]

Blogs are peer-reviewed too, just someone else's peers

Not sure swarm behaviour is a good idea, can a swarm move at the maximum speed of its members? In Spring I'd expect people to want their units to move as fast as possible.

[SwiftSpear]

Blogs are peer-reviewed too, just someone else's peers

Not sure swarm behaviour is a good idea, can a swarm move at the maximum speed of its members? In Spring I'd expect people to want their units to move as fast as possible.

You'd want the ability to set different swarm behaviors. Simple swarm navigation would just fix stupid things like units running into eachother and causing the fast unit to get stuck behind the slow unit and what not. Advanced usages could do all sorts of things like maintaining formations and modifying formations on the fly using efficient methods. Obviously we'd need appropriate UI and user control binds in place so it's still the player playing the game as opposed to the game playing for the player (or doing things the player doesn't want)

[edit] "peer reviewed reports" connotates a paper or article written by a really really smart person and then released specifically into the academic community. They describe things like how to make AI work at a mathematical level, or the detailed workings of a specific cancer cell, or an in depth analysis of 3 years of data collection in a psychological study. The fact that it's "peer reviewed" isn't really as important as the content attached to "peer reviewed reports" in general.

[DemO]

The main thing that pisses me off is when i make a drag move order and, for example:

Drag from top down somewhere on the map and the units decide to all cross past each other to carry out the order (unit at northmost point wants to go to southmost and vice versa)

Drag from bottom up and the units move nicely in the same line that they were before, or make a much better job of forming a line from a bunch.

I wish the orders would be the same regardless of what direction you drag from, because tbh i can't see ANY legitimate purpose for a line order that makes the units cross past and crash into each other for 5 seconds before forming a line.

[manored]

How about making it that then you add units to a group, they automatically become a swarm, doing things like moving at same speed and staying close to each other? Would be usefull for artillery & assault units coordination :)

[SwiftSpear]

Optimally, any unit in a selected group is treated as a swarm while in the duration of that selection. The problem is that it requires major pathfinding rewrites, which is in large part considered a black box in spring.