Difference between revisions of "InitialDesign"

From SUMOwiki
Jump to navigationJump to search
m (Initial Designs moved to Config:InitialDesign: <InitialDesign> tag links to this now)
 
 
(4 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
== Latin Hypercube Design ==
 
== Latin Hypercube Design ==
Choose an initial sampleset in such a way that they form a latin hypercube.
+
Choose an initial sampleset in such a way that they form a latin hypercube. This initial design will first try to download a pre-optimized maximin Latin hypercube from the website [http://www.spacefillingdesigns.nl]. If such a design is not available, or no proper connection to the website can be made, the Latin hypercube is generated by the toolbox. Note that this Latin hypercube may be suboptimal, and may leave considerable gaps in the design space.
 
{{OptionsHeader}}
 
{{OptionsHeader}}
 
{{Option
 
{{Option
Line 9: Line 9:
 
}}
 
}}
  
== Central Composite Design ==
+
== Dataset Design ==
 +
Reads an initial design from a dataset.  Each row should contain one sample.  The inputvalues come first.  The outputvalues, if there are any, come second.
  
{{OptionsHeader}}
+
If only input values are supplied (hasOutputs=false), each value has to be in the fixed [-1,1] range (model space). The samples are later automatically transformed to their original range before they are passed to the sample evaluator. This is done to easily support different ranges for variables without having to change the initial dataset design.
{{Option
 
|name        = type
 
|values     = [inscribed,factoral]
 
|default    = factoral
 
|description = Where are the '''''star points''''' placed? See http://www.itl.nist.gov/div898/handbook/pri/section3/pri3361.htm
 
}}
 
  
== Dataset Design ==
+
If, however, output values are also provided (hasOutputs=true), the samples are instantly fed to the toolbox for modelling, since they don't have to be evaluated anymore. Because of this property, the inputs have to be in the original range defined in the simulator xml file, instead of the fixed [-1,1] range.
Reads an initial design from a dataset. Each row should contain one sample.  The inputvalues come first.  The outputvalues, if there are any, come second.
 
  
 
{{OptionsHeader}}
 
{{OptionsHeader}}
Line 47: Line 41:
 
}}
 
}}
  
== Screening Designs ==
 
 
Create a design from information retrieved by parameter screening.
 
 
{{NodesHeader}}
 
{{Node
 
|name        = InitialDesign
 
|required    = yes
 
|description = Configuration for the requested InitialDesign Method.
 
}}
 
{{Node
 
|name        = ScreeningMethod
 
|required    = yes
 
|description = Configuration for the requested Screening Method.  See below for further instructions.
 
}}
 
 
=== ScreeningMethod: OneD Screening ===
 
Build one-dimensional metamodels (minimetamodel) for each factor.  Then choose the initial design based upon a factor importance metric.
 
{{OptionsHeader}}
 
{{Option
 
|name        = importanceMetric
 
|values      = [samples,sensitivity,minmax]
 
|default    = samples
 
|description = What defines the importance of a factor?  '''(1) samples:''' amount of samples required to reach a certain accuray.  '''(2) sensitivity:''' order of the metamodel, only works with polynomial models.  '''(3) minmax:''' greatest difference in the metamodel
 
}}
 
{{Option
 
|name        = configFile
 
|values      = path
 
|default    = '/config/defaultMini.xml'
 
|description = Path of the file that contains the configuration to create each minimetalmodel.  Best to leave this default. 
 
}}
 
  
 
== Factorial Design ==
 
== Factorial Design ==

Latest revision as of 15:41, 2 October 2009

Latin Hypercube Design

Choose an initial sampleset in such a way that they form a latin hypercube. This initial design will first try to download a pre-optimized maximin Latin hypercube from the website [1]. If such a design is not available, or no proper connection to the website can be made, the Latin hypercube is generated by the toolbox. Note that this Latin hypercube may be suboptimal, and may leave considerable gaps in the design space. Template:OptionsHeader Template:Option

Dataset Design

Reads an initial design from a dataset. Each row should contain one sample. The inputvalues come first. The outputvalues, if there are any, come second.

If only input values are supplied (hasOutputs=false), each value has to be in the fixed [-1,1] range (model space). The samples are later automatically transformed to their original range before they are passed to the sample evaluator. This is done to easily support different ranges for variables without having to change the initial dataset design.

If, however, output values are also provided (hasOutputs=true), the samples are instantly fed to the toolbox for modelling, since they don't have to be evaluated anymore. Because of this property, the inputs have to be in the original range defined in the simulator xml file, instead of the fixed [-1,1] range.

Template:OptionsHeader Template:Option Template:Option

Combined Design

Combines two other initial designs.

Template:NodesHeader Template:Node


Factorial Design

Choose an initial sampleset in such a way that they form an uniform grid Template:OptionsHeader Template:Option