class_path_and_name

Specify the module and class name of the python surrogate

Specification

• Alias: None

• Arguments: STRING

Description

Specify the module and class name of the python surrogate. At minimum, the class must contain the methods ‘construct’ and ‘predict’. Dakota passes to the former a matrix (num samples x num variables) of variable values and a num samples x 1 matrix of responses to use to train the surrogate. The predict method receives a matrix of num samples x num variables and expects a matrix of num samples x 1 predicted values to be returned. An optional ‘gradient’ method is supported as well and returns gradients for a matrix of variables passed from Dakota.

Examples

Use of python-based surrogates in Dakota is enabled via:

model,
    id_model = 'SURR'
    surrogate global,
    dace_method_pointer = 'DACE'
    experimental_python
        class_path_and_name = "surrogate_polynomial.Surrogate"

An corresponding python class (incomplete) would contain at minimum:

class Surrogate:

  def construct(self, var, resp):
      var2 = np.hstack((np.ones((var.shape[0], 1)), var))
      self.coeffs = np.zeros(var.shape[1])
      z = np.linalg.inv(np.dot(var2.T, var2))
      self.coeffs = np.dot(z, np.dot(var2.T, resp))
      return

  def predict(self, pts):
      return self.coeffs[0]+pts.dot(self.coeffs[1:])