Dilation (morphology): Difference between revisions

From formulasearchengine
Jump to navigation Jump to search
en>Omnipaedista
typo
 
Line 1: Line 1:
Hi there, I am Alyson Pomerleau and I believe it seems fairly good when you say it. She functions as a travel agent but soon she'll be on her own. Ohio is where my house is but my spouse desires us to transfer. As a woman what she truly likes is fashion and she's been doing it for fairly a whilst.<br><br>My web page: free psychic ([http://video.dd-club.org/uprofile.php?UID=26546 video.dd-club.org])
'''Sparse grids''' are numerical techniques to represent, integrate or interpolate high [[dimension]]al functions. They were originally developed by the [[Russia]]n [[mathematician]] [[Sergey A. Smolyak]], a student of [[Lazar Lyusternik]], and are based on a sparse tensor product construction. Computer algorithms for efficient implementations of such grids were later developed by [[Michael Griebel]] and [[Christoph Zenger]].
 
== [[Curse of dimensionality]] ==
The standard way of representing multidimensional functions are tensor or full grids. The number of basis functions or nodes (grid points) that have to be stored and processed [[exponential function|depend exponentially]] on the number of dimensions. Even with today's computational power it is not possible to process functions with more than 4 or 5 dimensions.
 
The curse of dimension is expressed in the order of the integration error that is made by a quadrature of level <math>l</math>, with <math>N_{l}</math> points. The function has regularity <math>r</math>, i.e. is <math>r</math> times differentiable. The number of dimensions is <math>d</math>.
 
<math>|E_l| = O(N_l^{-\frac{r}{d}})</math>
 
== Smolyak's quadrature rule ==
Smolyak found a computationally more efficient method of integrating multidimensional functions based on a univariate quadrature rule <math>Q^{(1)}</math>. The <math>d</math>-dimensional Smolyak integral <math>Q^{(d)}</math>of a function <math>f</math> can be written as a recursion formula with the tensor product.
 
<math>Q_l^{(d)} f = \left(\sum_{i=1}^l \left(Q_i^{(1)}-Q_{i-1}^{(1)}\right)\otimes Q_{l-i+1}^{(d-1)}\right)f</math>
 
The index to <math>Q</math> is the level of the discretization. A <math>1-d</math> integration on level <math>i</math> is computed by the evaluation of <math>O(2^{i})</math> points. The error estimate for a function of regularity <math>r</math> is:
 
<math>|E_l| = O\left(N_l^{-r}\left(\log N_l\right)^{(d-1)(r+1)}\right)</math>
 
 
== References ==
* [http://www.lrr.in.tum.de/~murarasu/ppopp027s-murarasu.pdf A memory efficient data structure for regular sparse grids]
* [http://sparse-grids.de/ Code to generate (and pre-generated) nodes and weights for quadrature]
* [http://wissrech.iam.uni-bonn.de/research/projects/zumbusch/fd.html Finite difference scheme on sparse grids]
* [http://www.vis.uni-stuttgart.de/ger/research/fields/recent/sparse/ Visualization on sparse grids]
* [http://wissrech.iam.uni-bonn.de/research/pub/garcke/kdd.pdf Datamining on sparse grids,  J.Garcke, M.Griebel (pdf)]
* [http://www.math.tu-berlin.de/~garcke/paper/sparseGridTutorial.pdf Sparse Grid Tutorial, J.Garcke (pdf)]
{{Mathanalysis-stub}}
[[Category:Numerical analysis]]

Revision as of 15:14, 4 December 2013

Sparse grids are numerical techniques to represent, integrate or interpolate high dimensional functions. They were originally developed by the Russian mathematician Sergey A. Smolyak, a student of Lazar Lyusternik, and are based on a sparse tensor product construction. Computer algorithms for efficient implementations of such grids were later developed by Michael Griebel and Christoph Zenger.

Curse of dimensionality

The standard way of representing multidimensional functions are tensor or full grids. The number of basis functions or nodes (grid points) that have to be stored and processed depend exponentially on the number of dimensions. Even with today's computational power it is not possible to process functions with more than 4 or 5 dimensions.

The curse of dimension is expressed in the order of the integration error that is made by a quadrature of level l, with Nl points. The function has regularity r, i.e. is r times differentiable. The number of dimensions is d.

|El|=O(Nlrd)

Smolyak's quadrature rule

Smolyak found a computationally more efficient method of integrating multidimensional functions based on a univariate quadrature rule Q(1). The d-dimensional Smolyak integral Q(d)of a function f can be written as a recursion formula with the tensor product.

Ql(d)f=(i=1l(Qi(1)Qi1(1))Qli+1(d1))f

The index to Q is the level of the discretization. A 1d integration on level i is computed by the evaluation of O(2i) points. The error estimate for a function of regularity r is:

|El|=O(Nlr(logNl)(d1)(r+1))


References

Template:Mathanalysis-stub