Sobol space coverage

Example to show the space coverage of Sobol-points on a unit-square.

Firstly we will create an simple Modelica model with two parameters like below:

model SobolExample
 parameter Real a;
 parameter Real b;
end SobolExample;

Save the SobolExample model to a file and upload it:

from modelon.impact.client import Client

client = Client(url=<impact-domain>) # url is optional; defaults to on-prem OR "https://impact.modelon.cloud"
workspace = client.create_workspace(<workspace-name>)

project = workspace.get_default_project()

project.import_modelica_library('SobolExample.mo').wait()

Next, we will create an experiment with Sobol expansion and 2 modifiers that are uniformly distributed in [0, 1) as follows:

from modelon.impact.client import Sobol, Uniform

# Choose analysis type
dynamic = workspace.get_custom_function('dynamic')

# Get model
model = workspace.get_model(<path to SobolExample model>)

# Set-up experiment
experiment_definition=experiment_definition.with_modifiers(
   {'a': Uniform(0, 1), 'b': Uniform(0, 1)}
   ).with_expansion(Sobol(50))

exp = workspace.create_experiment(experiment_definition)

To extract Sobol points, we can execute the experiment by calling the the execute method on the experiment object with the with_cases argument set to an empty list.This generate all the cases for a given experiment definition and compiles all FMUs needed for running the cases.

Note: No cases are executed when the execute method is called with an empty list argument as shown below:

experiment_expand = exp.execute(with_cases=[]).wait()

Now we can extract the Sobol points like below:

sobol_points = [list(case.input.parametrization.values()) for case in experiment_expand.get_cases()]

We can plot the points as shown below:

import matplotlib.pylab as pl

pl.figure()
pl.title(f"{50} Sobol points on unit square")
pl.scatter(*zip(*sobol_points), marker = '*', s = 48, color = 'blue')
pl.grid()
pl.xlim([0, 1])
pl.ylim([0, 1])
pl.show()