An Efficient Swarm Neighborhood Management for a 3D Tactical Simulator

The paper presents a unique utilization of the dynamic Delaunay triangulation (DT) to devise a highly efficient algorithm for the neighborhood adjacency management which is a crucial component of a swarm-based simulation. This algorithm allows fast computation of the swarm neighborhood which is required to implement Boids flocking rules. The method also provides an efficient mechanism to detect and manage object collisions. This unique utilization of DT based data structures was applied to a new 3D tactical swarm simulation called the Battle Swarm. The result was a very high speed of simulation, which made complex application of genetic evolution possible for goal-based swarms in a three dimensional space. The method´s effective and complex emergent behavior was successfully demonstrated by the Battle Swarm Simulator. In addition, the experimental section confirms that the method was highly efficient and can be used in similar applications.