Constrained Scikit-Optimize

About

Constrained Scikit-Optimize skopt_modcn is a modified version of Scikit-Optimize skopt. You can now add space_constraint to skopt.

Releases in pypi

PYPI Releases

Prerequisites

Constrained scikit-optimize requires

Python >= 3.6
NumPy (>= 1.13.3)
SciPy (>= 0.19.1)
joblib (>= 0.11)
scikit-learn >= 0.20
matplotlib >= 2.0.0
pydoe
scikit-optimize

How to install?

You can install the latest release with the following commands:

Step 1:

pip install scikit-optimize

Step 2:

pip install pydoe

Step 3:

pip install -i https://test.pypi.org/simple/ skopt-modcn

This installs the latest version of scikit-optimize-constrained.

How to use?

Then use the following in python:

import skopt_modcn

Sample optimization without constraints using the original `skopt`:

# Import skopt and numpy
import skopt
import numpy as np

# Define the objective function with two inputs
def f(x):
    # You can modify this function according to your problem
    return ( x[0] ) ** 2 + ( x[1] ) ** 2

# Define the domain of the inputs as two real intervals
space = [(-2.0, 2.0), (-2.0, 2.0)]

# Use skopt.gp_minimize function to find the minimum of the function
res = skopt.gp_minimize(f, space, n_calls=20, n_initial_points=10, initial_point_generator = "lhs",\
                        random_state=1, verbose=True)

# Get the best point and the best value observed so far
x_best = res.x
y_best = res.fun

print('The input of the minimum objective function is: '+ str(x_best))
print('The minimum of the objective function is: '+ str(y_best))

Sample optimization with constraints using `skopt_modcn`:

# Import skopt_modcn and numpy
import skopt_modcn
import numpy as np

# Define the objective function with two inputs
def f(x):
    # You can modify this function according to your problem
    return ( x[0] ) ** 2 + ( x[1] ) ** 2

def constraint(x):
    return (x[0] / x[1] <= 1) and (x[0] / x[1] >= 0)

# Define the domain of the inputs as two real intervals
space = [(-2.0, 2.0), (-2.0, 2.0)]

# Use skopt_modcn.gp_minimize function to find the minimum of the function
res = skopt_modcn.gp_minimize(f, space, n_calls=20, n_initial_points=10, initial_point_generator = "lhs_pyDOE",\
                              random_state=1, space_constraint=constraint, verbose=True)

# Get the best point and the best value observed so far
x_best = res.x
y_best = res.fun

print('The input of the minimum objective function is: '+ str(x_best))
print('The minimum of the objective function is: '+ str(y_best))

Additional features

Now, you can add the following options to use constrained initial points with your constrained optimization skopt_modcn:

initial_point_generator="grid_modified",

initial_point_generator="lhs_modified",

initial_point_generator="lhs_pydoe",

initial_point_generator="sobol_scipy",

Contributors

Developer of `skopt_modcn`

SaeednHT

Citation information

skopt_modcn was introduced in the following published paper:

Torbati, S., Daneshmehr, A., Pouraliakbar, H., Asgharian, M., Ahmadi Tafti, S. H., Shum-Tim, D., et al. (2024). Personalized evaluation of the passive myocardium in ischemic cardiomyopathy via computational modeling using Bayesian optimization. Biomech. Model. Mechanobiol. doi: 10.1007/s10237-024-01856-0 [https://doi.org/10.1007/s10237-024-01856-0]

About the original Scikit-Optimize:

Scikit-Optimize

Scikit-Optimize, or skopt, is a simple and efficient library to minimize (very) expensive and noisy black-box functions. It implements several methods for sequential model-based optimization. skopt aims to be accessible and easy to use in many contexts.

The library is built on top of NumPy, SciPy, and Scikit-Learn.

Scikit-Optimize does not perform gradient-based optimization. For gradient-based optimization algorithms please see

scipy.optimize

Important links

Skopt documentation - https://scikit-optimize.github.io/
Issue tracker - https://github.com/scikit-optimize/scikit-optimize/issues

SaeednHT/scikit-optimize-constrained