DOE algorithms options¶
A simple way to solve an OptimizationProblem
is to use the API method
execute_algo()
. E.g.
from gemseo.api import execute_algo
from gemseo.problems.analytical.rosenbrock import Rosenbrock
problem = Rosenbrock()
sol = execute_algo(problem, "OT_LHS", n_samples=20)
They are also used in all MDO and DOE scenarios in the dictionary
passed to the MDODiscipline.execute()
method.
List of available algorithms : CustomDOE - DiagonalDOE - OT_AXIAL - OT_COMPOSITE - OT_FACTORIAL - OT_FAURE - OT_FULLFACT - OT_HALTON - OT_HASELGROVE - OT_LHS - OT_LHSC - OT_MONTE_CARLO - OT_OPT_LHS - OT_RANDOM - OT_REVERSE_HALTON - OT_SOBOL - OT_SOBOL_INDICES - bbdesign - ccdesign - ff2n - fullfact - lhs - pbdesign -
CustomDOE¶
Description¶
This samples are provided either as a file in text or csv format or as a sequence of sequences of numbers.
Options¶
comments,
str
- The characters or list of characters used to indicate the start of a comment. None implies no comments.delimiter,
str
- The string used to separate values. If None, use whitespace.doe_file,
str
- Either the file, the filename, or the generator to read.eval_jac,
bool
- evaluate jacobianmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processessamples,
array
- The samples.skiprows,
int
- skip the first skiprows lineswait_time_between_samples,
float
- waiting time between two samples
DiagonalDOE¶
Description¶
Diagonal design of experiments
Options¶
eval_jac,
bool
- evaluate jacobianmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processesn_samples,
int
- number of samplesreverse,
list(str)
- list of dimensions or variables to sample from their upper bounds to their lower bounds. Default: None.wait_time_between_samples,
float
- waiting time between two samples
OT_AXIAL¶
Description¶
Axial design implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_COMPOSITE¶
Description¶
Composite design implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_FACTORIAL¶
Description¶
Factorial design implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_FAURE¶
Description¶
Faure sequence implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_FULLFACT¶
Description¶
Full factorial design implementedin openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_HALTON¶
Description¶
Halton sequence implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_HASELGROVE¶
Description¶
Haselgrove sequence implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_LHS¶
Description¶
Latin Hypercube Sampling implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_LHSC¶
Description¶
Centered Latin Hypercube Sampling implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_MONTE_CARLO¶
Description¶
Monte Carlo sequence implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_OPT_LHS¶
Description¶
Optimal Latin Hypercube Sampling implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_RANDOM¶
Description¶
Random sampling implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_REVERSE_HALTON¶
Description¶
Reverse Halton sequence implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_SOBOL¶
Description¶
Sobol sequence implemented in openTURNS library
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
OT_SOBOL_INDICES¶
Description¶
Sobol indices
External link¶
Options¶
annealing,
bool
- if True, use simulated annealing to optimize LHS. Otherwise, use crude Monte Carlo. Default: True.centers,
array
- centers for axial, factorial and composite designscriterion,
str
- space-filling criterion, either “C2”, “PhiP” or “MinDist”. Default: “C2”.distribution_name,
str
- distribution nameend,
float
- level end for trapezoidal distributioneval_jac,
bool
- evaluate jacobianlevels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)mu,
float
- mean of a random variable for beta, normal and truncated normal distributionsn_processes,
int
- number of processesn_replicates,
int
- number of Monte Carlo replicates to optimize LHS. Default: 1000.n_samples,
int
- number of samplesseed,
int
- seed value.sigma,
float
- standard deviation for beta, normal and truncated normal distributionsstart,
float
- level start for trapezoidal distributiontemperature,
string
- temperature profil for simulated annealing, either “Geometric” or “Linear”. Default: “Geometric”.wait_time_between_samples,
float
- waiting time between two samples
bbdesign¶
Description¶
Box-Behnken design implemented in pyDOE
External link¶
Options¶
alpha,
str
- effect the variance, either “orthogonal” or “rotatable”center_bb,
int
- number of center points for Box-Behnken designcenter_cc,
tuple
- 2-tuple of center points for the central composite designcriterion,
string
- Default value = None) :type criterion:eval_jac,
bool
- evaluate jacobianface,
str
- The relation between the start points and the corner (factorial) points, either “circumscribed”, “inscribed” or “faced”iterations,
integer
- Default value = 5) :type iterations:levels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processesn_samples,
int
- number of samplesseed,
int
- seed value.wait_time_between_samples,
float
- waiting time between two samples
ccdesign¶
Description¶
Central Composite implemented in pyDOE
External link¶
Options¶
alpha,
str
- effect the variance, either “orthogonal” or “rotatable”center_bb,
int
- number of center points for Box-Behnken designcenter_cc,
tuple
- 2-tuple of center points for the central composite designcriterion,
string
- Default value = None) :type criterion:eval_jac,
bool
- evaluate jacobianface,
str
- The relation between the start points and the corner (factorial) points, either “circumscribed”, “inscribed” or “faced”iterations,
integer
- Default value = 5) :type iterations:levels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processesn_samples,
int
- number of samplesseed,
int
- seed value.wait_time_between_samples,
float
- waiting time between two samples
ff2n¶
Description¶
2-Level Full-Factorial implemented in pyDOE
External link¶
Options¶
alpha,
str
- effect the variance, either “orthogonal” or “rotatable”center_bb,
int
- number of center points for Box-Behnken designcenter_cc,
tuple
- 2-tuple of center points for the central composite designcriterion,
string
- Default value = None) :type criterion:eval_jac,
bool
- evaluate jacobianface,
str
- The relation between the start points and the corner (factorial) points, either “circumscribed”, “inscribed” or “faced”iterations,
integer
- Default value = 5) :type iterations:levels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processesn_samples,
int
- number of samplesseed,
int
- seed value.wait_time_between_samples,
float
- waiting time between two samples
fullfact¶
Description¶
Full-Factorial implemented in pyDOE
External link¶
Options¶
alpha,
str
- effect the variance, either “orthogonal” or “rotatable”center_bb,
int
- number of center points for Box-Behnken designcenter_cc,
tuple
- 2-tuple of center points for the central composite designcriterion,
string
- Default value = None) :type criterion:eval_jac,
bool
- evaluate jacobianface,
str
- The relation between the start points and the corner (factorial) points, either “circumscribed”, “inscribed” or “faced”iterations,
integer
- Default value = 5) :type iterations:levels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processesn_samples,
int
- number of samplesseed,
int
- seed value.wait_time_between_samples,
float
- waiting time between two samples
lhs¶
Description¶
Latin Hypercube Sampling implemented in pyDOE
External link¶
Options¶
alpha,
str
- effect the variance, either “orthogonal” or “rotatable”center_bb,
int
- number of center points for Box-Behnken designcenter_cc,
tuple
- 2-tuple of center points for the central composite designcriterion,
string
- Default value = None) :type criterion:eval_jac,
bool
- evaluate jacobianface,
str
- The relation between the start points and the corner (factorial) points, either “circumscribed”, “inscribed” or “faced”iterations,
integer
- Default value = 5) :type iterations:levels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processesn_samples,
int
- number of samplesseed,
int
- seed value.wait_time_between_samples,
float
- waiting time between two samples
pbdesign¶
Description¶
Plackett-Burman design implemented in pyDOE
External link¶
Options¶
alpha,
str
- effect the variance, either “orthogonal” or “rotatable”center_bb,
int
- number of center points for Box-Behnken designcenter_cc,
tuple
- 2-tuple of center points for the central composite designcriterion,
string
- Default value = None) :type criterion:eval_jac,
bool
- evaluate jacobianface,
str
- The relation between the start points and the corner (factorial) points, either “circumscribed”, “inscribed” or “faced”iterations,
integer
- Default value = 5) :type iterations:levels,
array
- levels for axial, factorial and composite designsmax_time,
float
- maximum runtime in seconds, disabled if 0 (Default value = 0)n_processes,
int
- number of processesn_samples,
int
- number of samplesseed,
int
- seed value.wait_time_between_samples,
float
- waiting time between two samples