Home // ICN 2014, The Thirteenth International Conference on Networks // View article

A Flexible P2P Gossip-based PSO Algorithm

Authors:
Marco Biazzini

Keywords: P2P function optimization; distributed Particle Swarm Optimization; P2P distributed computation.

Abstract:
It is becoming more and more interesting, in the domain of distributed function optimization, the study of fully decentralized optimization algorithms, deployed on large networks of heterogeneous computational units. Several issues arise on such a system design, among which the proper way of distributing and making use of shared information, in absence of a centralized coordination, is a prominent one. We introduce the design of a P2P gossip-based Particle Swarm Optimization (PSO) algorithm, that is capable to implement different policies with respect to the use of global information, as this becomes available via gossiping during the computation. Such a PSO flavor is easy to tune, in order to implement different strategies, while balancing exploration and exploitation. Preliminary experimental results are shown to assess the usefulness of the proposal.

Pages: 81 to 85

Copyright: Copyright (c) IARIA, 2014

Publication date: February 23, 2014

Published in: conference

ISSN: 2308-4413

ISBN: 978-1-61208-318-6

Location: Nice, France

Dates: from February 23, 2014 to February 27, 2014