Particle swarms for dynamic optimization problems

Blackwell, T, Branke, J and Li, X 2008, 'Particle swarms for dynamic optimization problems' in C. Blum and D. Merkle (ed.) Swarm Intelligence: Introduction and Applications, Springer, Berlin, Germany, pp. 193-217.

Document type: Book Chapter
Collection: Book Chapters

Title Particle swarms for dynamic optimization problems
Author(s) Blackwell, T
Branke, J
Li, X
Year 2008
Title of book Swarm Intelligence: Introduction and Applications
Publisher Springer
Place of publication Berlin, Germany
Editor(s) C. Blum
D. Merkle
Start page 193
End page 217
Subjects Neural, Evolutionary and Fuzzy Computation
Summary Many practical optimization problems are dynamic in the sense that the best solution changes in time. An optimization algorithm, therefore, has to both find and subsequently track the changing optimum. examples include the arrival of new jobs in scheduling, changing expected profits in portfolio optimization, and fluctuating demand. Clearly, if the changes in the problem instance are radical, the best one can do is to repeatedly solve the optimization problem from scratch. However, in most practical applications the changes are gradual. If this is the case, it should be possible to speed up optimization after a problem change by utilizing some of the information on the fitness landscape gathered during the optimization process so far. In recent years, appropriately modified evolutionary algorithms (EAs) have been shown to achieve this successfully; see, e.g., [11, 23]; the focus of this chapter is to present similar advances within the context of particle swarm optimization. Particle swarm optimization (PSO) is a versatile population-based optimization technique, in many respects similar to evolutionary algorithms. Basically, particles 'fly' above the fitness landscape, while a particle's movement is influenced by its attraction to its neighborhood best (the best solution found by members of the particle's social network), and its personal best (the best solution the particle has found so far). PSO has been shown to perform well for many static problems [32], and is introduced in more detail in Section 2.
Copyright notice © 2008 Springer
Keyword(s) artificial intelligence
distributed systems
operations research
DOI - identifier 10.1007/978-3-540-74089-6_6
ISBN 9783540740889
Version Filter Type
Altmetric details:
Access Statistics: 2370 Abstract Views  -  Detailed Statistics
Created: Fri, 05 Mar 2010, 02:22:00 EST by Catalyst Administrator
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us