MAKE A MEME
View Large Image
| View Original: | PSO Meta-Fitness Landscape (12 benchmark problems).JPG (1042x689) | |||
| Download: | Original | Medium | Small | Thumb |
| Courtesy of: | commons.wikimedia.org | More Like This | ||
| Keywords: PSO Meta-Fitness Landscape (12 benchmark problems).JPG en Performance landscape showing how a simple Particle Swarm Optimization PSO variant performs in aggregate on several benchmark problems when varying two PSO parameters Lower meta-fitness values means better PSO performance Such a performance landscape is very time-consuming to compute especially for optimizers with several behavioural parameters but it can be searched efficiently using the simple meta-optimization approach by Pedersen implemented in http //www hvass-labs org/projects/swarmops/ SwarmOps to uncover PSO parameters with good performance Good choices would here seem to be in the region <math>\omega \in -0 6; 0</math> and <math>\phi_g \in 2; 4</math> and the region <math>\omega \in -1 25; -0 75</math> and <math>\phi_g \in -3 5; -2</math> Own Pedersen M E H http //www hvass-labs org/people/magnus/thesis/pedersen08thesis pdf Tuning Simplifying Heuristical Optimization PhD Thesis 2010 University of Southampton School of Engineering Sciences Computational Engineering and Design Group 2010-01-01 Evolutionary algorithms Optimization Benchmarks | ||||
.JPG){kind=link}
