SED:SED toolbox - Revision history

Tdhaene: /* Introduction */

2011-02-24T16:54:49Z

Introduction

← Older revision		Revision as of 16:54, 24 February 2011
Line 1:		Line 1:
	== Introduction ==		== Introduction ==
	[[Image:SED.png\|~~150~~ px\|right\|SED Toolbox]]		[[Image:SED.png\|110 px\|right\|SED Toolbox]]

	The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ''Design of Experiments (DoE)''.		The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ''Design of Experiments (DoE)''.

Kcrombec: /* Introduction */

2011-02-24T13:38:44Z

Introduction

← Older revision		Revision as of 13:38, 24 February 2011
Line 1:		Line 1:
	== Introduction ==		== Introduction ==
	[[Image:SED.~~gif~~\|~~250~~ px\|right\|SED Toolbox]]		[[Image:SED.png\|150 px\|right\|SED Toolbox]]

	The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ''Design of Experiments (DoE)''.		The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ''Design of Experiments (DoE)''.

Kcrombec: /* Introduction */

2011-01-26T15:31:46Z

Introduction

← Older revision		Revision as of 15:31, 26 January 2011
Line 11:		Line 11:
	The ''intersite distance'' is the smallest distance between two design points in the design space; this value should be as high as possible, in order to have the points spread out as evenly as possible.		The ''intersite distance'' is the smallest distance between two design points in the design space; this value should be as high as possible, in order to have the points spread out as evenly as possible.
	In addition to the intersite distance, the projected distance is also important.		In addition to the intersite distance, the projected distance is also important.
	The ''projected distance'' is the smallest distance between all the points after they have been projected on one ~~axis~~ of the design space. This measure is especially important if the relative importance of the design parameters is unknown. E.g., if one of the design parameters does not influence the output behavior, two design points which only differ in this (irrelevant) parameter have the same behavior, and can be seen as the "same" design point. Thus, the projected distance must also be maximized.		The ''projected distance'' is the smallest distance between all the points after they have been projected on one axes of the design space. This measure is especially important if the relative importance of the design parameters is unknown. E.g., if one of the design parameters does not influence the output behavior, two design points which only differ in this (irrelevant) parameter have the same behavior, and can be seen as the "same" design point. Thus, the projected distance must also be maximized.

	All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.		All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.

Kcrombec: /* Quick start guide */

2011-01-26T13:16:36Z

Quick start guide

← Older revision		Revision as of 13:16, 26 January 2011
Line 28:		Line 28:
	</source>		</source>

	Now the toolbox is ready to be used. The SED Toolbox can be used in several ways, based on how much freedom you want in configuring and fine-tuning the parameters of the algorithms. We will now describe the three ways the toolbox can be used, in order of complexity, based on your requirements.		Now the toolbox is ready to be used. The SED Toolbox can be used in several ways, based on how much freedom you want in configuring and fine-tuning the parameters of the algorithms. We will now describe the three ways the toolbox can be used, in order of complexity, based on your requirements. If you prefer to learn by example, you can check out the examples directory in the distribution, which contains several applications and example problems for the toolbox.

	=== You want an ND design of X points ===		=== You want an ND design of X points ===

Kcrombec: /* TODO */

2011-01-24T14:22:57Z

TODO

← Older revision		Revision as of 14:22, 24 January 2011
Line 99:		Line 99:
	The Monte Carlo methods scale very well with the number of dimensions and points and should work for high-dimensional problems. However, the optimizer methods suffer more from the curse of dimensionality. '''optimizer-intersite''' should work up to 10D, but will run into memory problems for higher dimensions.		The Monte Carlo methods scale very well with the number of dimensions and points and should work for high-dimensional problems. However, the optimizer methods suffer more from the curse of dimensionality. '''optimizer-intersite''' should work up to 10D, but will run into memory problems for higher dimensions.

	~~== TODO ==~~
	~~TODO:~~
	* mc-intersite-projected-th en optimizer-intersite kunnen zonder kandidaten vallen
	* optimizer-projected unsupported
	* zie selectie van punten gebeuren met optimizer-intersite

	== Contribute ==		== Contribute ==

	Suggestions on how to improve the SED toolbox are always welcome. For more information please see the [[Feedback \| feedback]] page.		Suggestions on how to improve the SED toolbox are always welcome. For more information please see the [[Feedback \| feedback]] page.

Kcrombec: /* Constraints */

2011-01-24T13:34:06Z

Constraints

← Older revision		Revision as of 13:34, 24 January 2011
Line 87:		Line 87:
	=== Constraints ===		=== Constraints ===

	the default method '''mc-intersite-projected-threshold''' can run into problems when you are using very strict constraints. Because the Monte Carlo points are filtered by the projected distance threshold, it might be possible that no candidates remain that satisfy the constraints. In that case, '''mc-intersite-projected''' can be a good alternative. It produces slightly worse designs but is much more robust in terms of constraints.		the default method '''mc-intersite-projected-threshold''' can run into problems when you are using very strict constraints. Because the Monte Carlo points are filtered by the projected distance threshold, it might be possible that no candidates remain that satisfy the constraints. In that case, '''mc-intersite-projected''' can be a good alternative. It produces slightly worse designs but is much more robust in terms of constraints. Additionally, '''mc-intersite-projected-threshold''' and all other methods besides '''mc-intersite-projected''' need the corner points [-1,...,-1] and [1,...,1] to start, and if they violate the constraints they will still be selected. You can later request the design without these corner points using the getAllPointsWithoutInitialDesign() function, so this might not be an issue, but keep it in mind.

	=== Quality vs speed ===		=== Quality vs speed ===

Tdhaene: /* Introduction */

2011-01-23T17:45:54Z

Introduction

← Older revision		Revision as of 17:45, 23 January 2011
Line 11:		Line 11:
	The ''intersite distance'' is the smallest distance between two design points in the design space; this value should be as high as possible, in order to have the points spread out as evenly as possible.		The ''intersite distance'' is the smallest distance between two design points in the design space; this value should be as high as possible, in order to have the points spread out as evenly as possible.
	In addition to the intersite distance, the projected distance is also important.		In addition to the intersite distance, the projected distance is also important.
	The ''projected distance'' is the smallest distance between all the points after they have been projected on one axis of the design space. This measure is especially important if the relative importance of the design parameters is unknown. E.g., if one of the design parameters does not influence the output behavior, two design points which only differ in this (irrelevant) parameter have the same behavior, and can be seen as ~~one~~ design point. Thus, the projected distance must also be maximized.		The ''projected distance'' is the smallest distance between all the points after they have been projected on one axis of the design space. This measure is especially important if the relative importance of the design parameters is unknown. E.g., if one of the design parameters does not influence the output behavior, two design points which only differ in this (irrelevant) parameter have the same behavior, and can be seen as the "same" design point. Thus, the projected distance must also be maximized.

	All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.		All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.

Tdhaene: /* Introduction */

2011-01-23T17:44:33Z

Introduction

← Older revision		Revision as of 17:44, 23 January 2011
Line 3:		Line 3:

	The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ''Design of Experiments (DoE)''.		The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ''Design of Experiments (DoE)''.
	In traditional experimental design, all the design points are selected up front before performing any (computer or real-life) experiment, and afterwards no additional design points are selected. This traditional approach is prone to oversampling and/or undersampling, because it is often very difficult to estimate up front the required number of design points.		In traditional experimental design, all the design points are selected up front, before performing any (computer or real-life) experiment, and afterwards, no additional design points are selected. This traditional approach is prone to oversampling and/or undersampling, because it is often very difficult to estimate up front the required number of design points.

			The SED Toolbox solves this problem by providing the user with state-of-the-art algorithms that generate an experimental design in a sequential way, i.e. one design point at a time, without having to provide the total number of design points in advance. This is called ''sequential experimental design (SED)''. The SED Toolbox was designed to be extremely fast and easy to use, yet very powerful.

	The SED Toolbox solves this problem by providing the user with state-of-the-art algorithms that generate a design of experiments one point at a time, without having to provide the total number of design points in advance. This is called "sequential experimental design (SED)". The SED Toolbox was designed to be extremely fast and easy to use, yet very powerful.

	Central to the experimental design problem is the trade-off between the ''intersite'' (maximin) and ''projected'' (non-collapsing) requirements.		Central to the experimental design problem is the trade-off between the ''intersite'' (maximin) and ''projected'' (non-collapsing) requirements.
	The ''intersite distance'' is the smallest distance between two points in the design; this value should be as high as possible, in order to have the points spread out as evenly as possible.		The ''intersite distance'' is the smallest distance between two design points in the design space; this value should be as high as possible, in order to have the points spread out as evenly as possible.
	In addition to the intersite distance, the projected distance is also important.		In addition to the intersite distance, the projected distance is also important.
	The ''projected distance'' is the smallest distance between all the points after they ~~are~~ projected on one of the ~~axes~~. This is an important ~~property~~ if ~~it is unknown up front how important each~~ design parameters ~~included in the experiment~~ is. If one of the design parameters ~~is irrelevant~~, two points which differ ~~only~~ in this ~~value~~ can be ~~considered the same~~ point. Thus, the projected distance must also be maximized.		The ''projected distance'' is the smallest distance between all the points after they have been projected on one axis of the design space. This measure is especially important if the relative importance of the design parameters is unknown. E.g., if one of the design parameters does not influence the output behavior, two design points which only differ in this (irrelevant) parameter have the same behavior, and can be seen as one design point. Thus, the projected distance must also be maximized.

	All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.		All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.

Tdhaene at 16:38, 23 January 2011

2011-01-23T16:38:53Z

← Older revision		Revision as of 16:38, 23 January 2011
Line 1:		Line 1:
	== Introduction ==		== Introduction ==
	[[Image:~~blindDACE~~.gif\|250 px\|right\|SED Toolbox]]		[[Image:SED.gif\|250 px\|right\|SED Toolbox]]

	The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ~~design~~ of ~~experiments~~. In traditional design ~~of experiments~~, ~~the~~ all the design points are selected ~~at once~~, and ~~all the experiments are performed at once without selecting any~~ additional design points. This ~~method~~ is prone to ~~over-~~ or undersampling, because it is often very difficult to ~~predict~~ the required number of design points ~~in advance~~.		The SED Toolbox (Sequential Experimental Design) is a powerful Matlab toolbox for the sequential ''Design of Experiments (DoE)''.
			In traditional experimental design, all the design points are selected up front before performing any (computer or real-life) experiment, and afterwards no additional design points are selected. This traditional approach is prone to oversampling and/or undersampling, because it is often very difficult to estimate up front the required number of design points.

	The SED Toolbox solves this problem by providing the user with state-of-the-art algorithms that generate a design of experiments one point at a time, without having to provide the total number of design points in advance. This is called sequential experimental design. The SED Toolbox was designed to be extremely fast and easy to use, yet very powerful.		The SED Toolbox solves this problem by providing the user with state-of-the-art algorithms that generate a design of experiments one point at a time, without having to provide the total number of design points in advance. This is called "sequential experimental design (SED)". The SED Toolbox was designed to be extremely fast and easy to use, yet very powerful.

	Central to the experimental design problem is the trade-off between the intersite (maximin) and projected (non-collapsing) requirements. The intersite distance is the smallest distance between two points in the design; this value should be as high as possible, in order to have the points spread out as evenly as possible. In addition to the intersite distance, the projected distance is also important. The projected distance is the smallest distance between all the points after they are projected on one of the axes. This is an important property if it is unknown up front how important each design parameters included in the experiment is. If one of the design parameters is irrelevant, two points which differ only in this value can be considered the same point. Thus, the projected distance must also be maximized. All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.		Central to the experimental design problem is the trade-off between the ''intersite'' (maximin) and ''projected'' (non-collapsing) requirements.
			The ''intersite distance'' is the smallest distance between two points in the design; this value should be as high as possible, in order to have the points spread out as evenly as possible.
			In addition to the intersite distance, the projected distance is also important.
			The ''projected distance'' is the smallest distance between all the points after they are projected on one of the axes. This is an important property if it is unknown up front how important each design parameters included in the experiment is. If one of the design parameters is irrelevant, two points which differ only in this value can be considered the same point. Thus, the projected distance must also be maximized.

			All the algorithms in the SED Toolbox were optimized to produce designs that score well on both the intersite and projected distance.

	== Download ==		== Download ==

Kcrombec: /* TODO */

2011-01-20T16:36:46Z

TODO

← Older revision		Revision as of 16:36, 20 January 2011
Line 94:		Line 94:
	== TODO ==		== TODO ==
	TODO:		TODO:
	~~- problemen:~~		* mc-intersite-projected-th en optimizer-intersite kunnen zonder kandidaten vallen
	- mc-intersite-projected-th en optimizer-intersite kunnen zonder kandidaten vallen		* optimizer-projected unsupported
	- optimizer-projected unsupported		* zie selectie van punten gebeuren met optimizer-intersite
	- zie selectie van punten gebeuren met optimizer-intersite


	~~--TO BE UPDATED--~~

	two scripts dacefit.m and predictor.m that emulate the behavior of the DACE toolbox ([http://www2.imm.dtu.dk/~hbn/dace/]). Note, that full compatibility between blindDACE and the DACE toolbox is not provided. The scripts merely aim to ease the transition from the DACE toolbox to the blindDACE toolbox.

	~~Example code:~~
	~~<source lang="matlab">~~
	~~krige = dacefit(samples, values, 'regpoly0', 'corrgauss', theta0, lb, ub )~~
	~~y = predictor([1 2], krige)~~
	~~</source>~~

	Obviously, a lot less code is used to copy the setup described above. However, less code means less flexibility (e.g., blind kriging and regression kriging are not available using the wrapper scripts). Hence, it is suggested to learn the object oriented interface of SED and use it instead.

	== Contribute ==		== Contribute ==

	Suggestions on how to improve the SED toolbox are always welcome. For more information please see the [[Feedback \| feedback]] page.		Suggestions on how to improve the SED toolbox are always welcome. For more information please see the [[Feedback \| feedback]] page.