Extended Flocking Algorithm for Self-parameter Tuning

Abstract

Flocking algorithms for a multi-agent system are distributed algorithms that only have simple rules for each agent but generate complex formational movement. These algorithms are known as swarm intelligence and are robust and disaster tolerant for most cases. We consider that flocking algorithms that have these characteristics are the way to generate homeostasis in a system. We expect that by making use of this algorithm the system can tune its self parameters and thus maintain a high performance. First, to apply a flocking algorithm to a system, we extended the flocking algorithm to form an arbitrary lattice for further flexibility. We then applied extended flocking algorithm to position tracking camera system as an example.