Name
Title | ||
|---|---|---|
Decentralised distributed multiple objective particle swarm optimisation using peer to peer networks |
Abstract | ||
|---|---|---|
This paper describes a distributed particle swarm optimisation algorithm (PSO) based on peer-to-peer computer networks. A number of modifications are made to the more traditional synchronous PSO algorithm to allow for fully de- centralised, scalable and fault-tolerent operation. The modified algorithm uses staggered propagation of objective-space knowl- edge between sub-swarms to eliminate the need for a centralised data store. Analytical test functions are used to examine the performance of the proposed algorithm and its variations in comparison with a basic synchronous PSO implementation. The results clearly show the feasibility of decentralised particle swarm optimisation. I. INTRODUCTION Computational optimisation is becoming more and more prevalent in a number of fields where simple analytical solutions to a problem are not possible. Typically, intelligent search algorithms are used, which attempt to locate optimal solutions to a problem involving a number of different parameters through systematic evaluation of a subset of all possible solutions. For problems with many parameters and extremely large search spaces, utilising distributed computing resources becomes an attractive proposition, especially when individual solutions require significant computational time. Typically, specialised supercomputers or computer clusters are used for these applications, however increasing demand for such resources often means limited availability and long wait times. To address this problem, a number of systems have been proposed which bring together idle computing resources (desktop PCs, laptops, servers etc.) over a peer to peer network to form temporary ad-hoc computational grids (1). While these systems can significantly increase the computational power available, they pose a number of prob- lems to traditional distributed optimisation algorithms. This paper discusses these issues, and proposes a decentralised, distributed multiple objective particle swarm optimisation algorithm designed to operate within a peer-to-peer grid environment. |
Year | DOI | Venue |
|---|---|---|
2008 | 10.1109/CEC.2008.4631191 | Evolutionary Computation, 2008. CEC 2008. |
Keywords | Field | DocType |
multivariable systems,particle swarm optimisation,peer-to-peer computing,decentralised distributed multiple objective particle swarm optimisation,fault-tolerent operation,objective-space knowledge,peer-to-peer computer networks,staggered propagation | Particle swarm optimization,Approximation algorithm,Mathematical optimization,Algorithm design,Grid computing,Peer-to-peer,Computer science,Evolutionary computation,Genetic algorithm,Distributed computing,Scalability | Conference |
ISBN | Citations | PageRank |
978-1-4244-1823-7 | 9 | 0.67 |
References | Authors | |
6 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Scriven, I. | 1 | 20 | 1.38 |
| Andrew Lewis | 2 | 15 | 4.30 |
| David Ireland | 3 | 32 | 5.98 |
| Junwei Lu | 4 | 387 | 34.94 |