Statistics – Applications
Scientific paper
Dec 2002
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2002aas...201.9104e&link_type=abstract
American Astronomical Society, 201st AAS Meeting, #91.04; Bulletin of the American Astronomical Society, Vol. 34, p.1256
Statistics
Applications
Scientific paper
We report on progress in the development of general-purpose algorithms for global parameter minimization in scientific applications where comparison of results with data takes the form of a pattern recognition problem. Our basic approach is to implement model calculations for the problem of interest in parallel on a Beowulf cluster, with a genetic algorithm to optimize the parameters of the calculation and a neural network trained on observational data to compute the fitness function for members of the genetic population in each successive generation. In the specific application discussed in this presentation, we investigate galaxy collisions using a gravity tree plus SPH hydrodynamics (the Max Planck code GADGET), our own genetic algorithm code for global parameter minimization, and our own neural network code for comparison of calculations with observational data.
Edirisinghe D.
Edirisinghe G.
Guidry Mike
Messer O.
No associations
LandOfFree
Genetic Algorithms and Neural Networks Applied to Galaxy Collisions does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Genetic Algorithms and Neural Networks Applied to Galaxy Collisions, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Genetic Algorithms and Neural Networks Applied to Galaxy Collisions will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1890845