sciope.designs package

Submodules

sciope.designs.factorial_design module

Factorial Initial Design

class sciope.designs.factorial_design.FactorialDesign(levels, xmin, xmax, use_logger=False)

Bases: sciope.designs.initial_design_base.InitialDesignBase

Class definition for Factorial design

Properties/variables: * name (FactorialDesign) * levels (the number of levels in the factorial design) * xmin (lower bound of multi-dimensional space encompassing generated points) * xmax (upper bound of multi-dimensional space encompassing generated points) * outlier_column_indices (columns containing outliers) * logger (a logging object to display/save events) * use_logger (a boolean variable controlling whether logging is enabled or disabled)

Methods: * generate (returns a delayed object that can generated the desired number of samples)

draw(n_samples, chunk_size=1, auto_redesign=True)

Draw specified number of points from a generated LHD

n_samplesinteger

[]

n: integer

[]

auto_redesignboolean

[]

vector/array

generate()

Sub-classable method for generating a factorial design of specified ‘levels’ in the given domain. The number of generated points is levels^d.

dask.delayed

sciope.designs.initial_design_base module

Initial Design Base Class

class sciope.designs.initial_design_base.InitialDesignBase(name, xmin, xmax, use_logger=False)

Bases: object

Base class for initial designs. Must not be used directly! Each initial design type must implement the methods described herein:

Properties/variables: * name (FactorialDesign) * xmin (lower bound of multi-dimensional space encompassing generated points) * xmax (upper bound of multi-dimensional space encompassing generated points) * logger (a logging object to display/save events - set by derived classes) * use_logger (a boolean variable controlling whether logging is enabled or disabled)

Methods: * generate (returns the generated samples) * scale_variable (scales a variable from an old domain range to a new domain range) * scale_to_new_domain (scales a matrix from an old domain range to a new domain range)

abstract generate(n)

Sub-classable method for generating ‘n’ points within a given domain. Each derived class must implement.

static scale_to_new_domain(x, new_min, new_max)

Scales a given array/matrix to a new range

xmultidimensional array/matrix

to operate upon

new_minvector

the new lower bound of the domain, one element per dimension

new_maxvector

the new upper bound of the domain, one element per dimension

matrix

scaled to the new range

static scale_variable(x, old_min, old_max, new_min, new_max)

Scales a dimension from the specified old range to a new range.

xvector

represents a variable or dimension to operate upon

old_minvector

the old lower bound of the domain

old_maxvector

the old upper bound of the domain

new_minvector

the new lower bound of the domain

new_maxvector

the new upper bound of the domain

vector

scaled to the new range

sciope.designs.latin_hypercube_sampling module

Latin Hypercube Sampling Initial Design Implemented as a Translational Propagation LHD Implementation follows structure from the original reference below.

Ref: Viana, Felipe AC, Gerhard Venter, and Vladimir Balabanov. “An algorithm for fast optimal Latin hypercube design of experiments.” International journal for numerical methods in engineering 82, no. 2 (2010): 135-156.

class sciope.designs.latin_hypercube_sampling.LatinHypercube(xmin, xmax, use_logger=False, seed_size=None)

Bases: sciope.designs.initial_design_base.InitialDesignBase

Translational Propagation Latin Hypercube Sampling

Properties/variables: * name (LatinHypercube) * xmin (lower bound of multi-dimensional space encompassing generated points) * xmax (upper bound of multi-dimensional space encompassing generated points) * logger (a logging object to display/save events - set by derived classes) * use_logger (a boolean variable controlling whether logging is enabled or disabled) * seed_size (number of points in the LHS seed design)

Methods: * generate (returns the generated samples)

draw(n_samples, n=50, chunk_size=1, auto_redesign=True)

Draw specified number of points from a generated LHD

n_samplesinteger

[]

n: integer

[]

auto_redesignboolean

[]

vector/array

generate

generate_array(n, chunk_size=('auto', 'auto'))

Generate a partial design of specified points

n: integer

# of desired points

vector/array

sciope.designs.random_sampling module

Random Sampling Initial Design

class sciope.designs.random_sampling.RandomSampling(xmin, xmax, use_logger=False)

Bases: sciope.designs.initial_design_base.InitialDesignBase

Random Sampling implemented through gpflowopt

Properties/variables: * name (RandomSampling) * xmin (lower bound of multi-dimensional space encompassing generated points) * xmax (upper bound of multi-dimensional space encompassing generated points) * logger (a logging object to display/save events - set by derived classes) * use_logger (a boolean variable controlling whether logging is enabled or disabled)

Methods: * generate (returns the generated samples)

generate

Module contents