Asynchronous Particle Swarm Optimization for Swarm Robotics

Abstract

In the original particle swarm optimization algorithm, particles’ update is done synchronously. The whole swarm fitness is evaluated first before particle update process is conducted. Whereas in asynchronous update a particle is able to update its velocity and position after its fitness is evaluated. This caused the particle's search to be conducted with imperfect information. However, asynchronous update is useful in field such as swarm robotics search problem, where the robots can move continuously based on the available information without waiting for the whole swarm. Hence this paper looks into the differences between synchronous and asynchronous PSO and its application in swarm robotics search.