Iterations module

Describe the iterative process performed to find and optimized trajectory that satisfies the constraints.

pyrotor.iterations.binary_search_best_trajectory(trajectory, i, step, verbose)

Perform a binary search amoung all the kappas in order to find the optimised trajectory that verifies constraints.

Inputs:

• trajectory: Pyrotor instance

  Trajectory to optimize with associated attributes; the object will be directly modified by reference

• i: int

  Index of the kappa to use

• step: int

  Size of the current split

• verbose: boolean

  Display pr not the verbose

pyrotor.iterations.compute_kappa_max(kappa_mean, opti_factor)

Compute the supposed possible maximum value of kappa.

Inputs:

• kappa_mean: float

  Mean of kappa

• opti_factor: float

  Optimisation factor

Output:

• kappa_max: float

  Supposed possible maximum value of kappa

pyrotor.iterations.compute_kappa_mean(evaluations_f, evaluations_g)

Compute the mean of kappa.

Inputs:

• evaluations_f: list

  Evaluations of f over reference coefficients

• evaluations_g: list

  Evaluations of g over reference coefficients

Output:

kappa_mean: float

Mean of kappa

pyrotor.iterations.get_kappa_boundaries(reference_coefficients, format_model, sigma_inverse, c_weight, opti_factor, extra_info)

Give the possible minumum and maximum supposed values of kappa.

Here the cost function is of the form f + kappa*g so kappa is chosen in such a way that kappa = opti_factor * mean f / mean g where opti_factor is chosen by the user.

The larger opti_factor is, the more important g is when optimising.

Inputs:

• reference_coefficients: ndarray

  Coefficients of reference trajectories

• format_model: list of arrays or sklearn model

  Model of the cost; if list, the first element of the list is the the integrated linear part W and the second one the integrated quadratic part Q

• sigma_inverse: ndarray

  Pseudoinverse of the covariance matrix of the reference coefficients

• c_weight: ndarray

  Coefficients of a weighted trajectory

• opti_factor: float

  Optimisation factor

• extra_info: dict

  Contains independent_variable, basis, basis_features and basis_dimension dictionaries

Outputs:

• kappa_min: float

  Supposed possible minimum value of kappa; here set to 0

• kappa_max: float

  Supposed possible maximum value of kappa

pyrotor.iterations.iterate_through_kappas(trajectory, kappa_min, kappa_max, verbose)

Iterate through the different kappas in order to find the optimised trajectory that verifies constraints.

Inputs:

• trajectory: Pyrotor instance

  Trajectory to optimize with associated attributes; the object will be directly modified by reference

• kappa_min: float

  Supposed minimum possible kappa

• kappa_max: float

  Supposed maximum possible kappa

• verbose: boolean

  Display or not the verbose