rapid_models.doe

Basic DOE package for rapid-models based on pyDOE2.

Submodules

Package Contents

Functions

fullfact_with_bounds(LBs, UBs, N_xi)

Return a ND array of corresponding (x_0, ..., x_i) values that span out the inputspace

lhs_with_bounds(nDim, nSamples, LBs, UBs[, random_state])

Return a 2D array of corresponding (x, y) values that fill the input space

in_hull(p, hull)

Test if points in p are in hull

kmeans_sample(points, N[, values, mode, random_state])

Based on
rapid_models.doe.fullfact_with_bounds(LBs, UBs, N_xi)

Return a ND array of corresponding (x_0, …, x_i) values that span out the inputspace between lowerbound and upperbound in a structured (grid-like) way with n_x_i points in the i’th-dimension.

Parameters:

  • LBs (list-like, 1D) – lower bounds of the input space. len(LBs) must equal len(UBs)

  • UBs (list-like, 1D) – upper bounds of the input space. len(UBs) must equal len(LBs)

  • N_xi (list-like, 1D) – number of equidistant samples for each xi-dimension

Returns:

[[x_0,…,x_i],…,[x_0_n,…,x_i_n]]

Return type:

fullfact (ndarray)

rapid_models.doe.lhs_with_bounds(nDim, nSamples, LBs, UBs, random_state=None)

Return a 2D array of corresponding (x, y) values that fill the input space between lowerbound and upperbound with n points using a Latin-hypercube design.

Parameters:

  • nDim (int) – Number of dimensions

  • nSamples (int) – Number of total samples

  • LBs (list-like) – 1D, lower bounds of the input space. len(LBs) must equal len(UBs)

  • UBs (list-like) – 1D, upper bounds of the input space. len(UBs) must equal len(LBs)

  • random_state (int, RandomState instance or None, default=None) – Determines random number generation used to initialize the samples. Pass an int for reproducible results across multiple function calls.

Returns:

Array of sample points with shape (nSamples, nDim)[[x_0,…,x_i],…,[x_0_n,…,x_i_n]]

Return type:

lhs (ndarray)

rapid_models.doe.in_hull(p, hull)

Test if points in p are in hull

p should be a NxK coordinates of N points in K dimensions hull is either a scipy.spatial.Delaunay object or the MxK array of the coordinates of M points in `K`dimensions for which Delaunay triangulation will be computed.

Parameters:

  • p (array-like) – NxK array. Set of N points in K dimensions for which to check if is inside convex hull.

  • hull (array like / scipy.spatial.Delaunay) – MxK array. Set of M points in K dimensions to calculate the Delaunay tessellation using the [Qhull library](http://www.qhull.org/), or an existing scipy.spatial.Delaunay object.

Returns:

(N, ) boolean array of sample points where

True’ indicate that the point of that index is inside the triangulation while `False indicate that the point of corresponding index is outside the triangulation.

Return type:

b_in_hull (boolean ndarray)

rapid_models.doe.kmeans_sample(points, N, values=None, mode='center', random_state=42)

Based on https://stackoverflow.com/questions/69195903/

Parameters:

  • points (array-like) – PxK array, set of ´P´ points in ´K´ dimensions from which to calculate N k-means clusters to use as distance maximizing samples in K-dimensions

  • N (int) – Number of clusters

  • values (array-like, 1D) – Array of corresponding values at each point ´P´ in points. These values can be used to select e.g. minimum or maximum values inside each cluster by specifying mode

  • mode (str, default="center") – default mode select center of k-means cluster. “center_closest_point” select the point which are closest to the cluster center, “min” select the point with the corresponding minimum value, “max” select the point with the corresponding maximum value

Returns:

NxK array of ´N´ points in ´K´ dimensions

representing the mode (default “center”) in the N k-means clusters.

values (ndarray): optional array of N values of the returned points.

Return type:

points (ndarray)