Back to General and Gameplay Programming

Finding neighbouring agents within interaction range efficiently

dangerdaveCS · 2010-11-02T14:39:21

Hello all. I have a bunch of critters that interact with each other only when nearby (i.e. they communicate by touch). So given an agent, I need to find all other agents within interaction range.My first intuition is boids-style bin lattice spatial subdivision (i.e. divide all of space into a grid, each cell storing pointers to agents).However, I've recently been playing around with Bullet physics library, and notice that uses a completely different method(s) for its broadphase stage of collision detection (which is basically what I'm trying to do).I must not be understanding something right, because I can't think of a more efficient data structure than a grid, yet a popular physics library uses some mad hierarchical axis aligned something or other.Can someone please enlighten me?

General and Gameplay Programming Programming

Started by dangerdaveCS October 28, 2010 06:35 AM

9 comments, last by Adam_42 13 years, 5 months ago

Adam_42

3,664

November 02, 2010 02:39 PM

Testing each algorithm is the only way to be certain which one is quickest for your use.

However A 2D grid is usually fast enough if the size of the grid cells is chosen well. In 3D however grids can be very wasteful of memory.

Finding neighbouring agents within interaction range efficiently

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Finding neighbouring agents within interaction range efficiently

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines