Title: On the optimal size of a robust swarm

Authors: Mun-Loong Kam

Addresses: Cooperative Systems Lab (CoSyL), Department of Mechanical Engineering, National University of Singapore, Engineering Drive 2, 117576 Singapore

Abstract: A systematic approach to optimising swarm size is proposed to ensure the robustness when individual swarm units possibly fail while the interference among the units is maintained at an acceptable level. Considering failures of swarm units (e.g. motor and control system) or a swarm system deployed for a hazardous mission, each swarm unit has a certain failure probability. An increased swarm size will guard the system against failure although incurring a cost – a greater interference among the swarm units which will hinder the overall swarm performance. This work intends to extend prior work done by various researchers to optimise the swarm size by combining the above-mentioned effects of increasing swarm size. Models for the benefit and cost of increasing swarm size are presented and used to solve this optimisation problem for a robust swarm.

Keywords: robotic swarms; robustness; optimisation; optimal swarm size; failure probability.

DOI: 10.1504/IJCAT.2011.042691

International Journal of Computer Applications in Technology, 2011 Vol.41 No.3/4, pp.160 - 168

Published online: 26 Sep 2011 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article