DiscreteModel#

class DiscreteModel(operators)[source]#

Nonparametric discrete dynamical system model \(\qhat_{j+1} = \fhat(\qhat_{j}, \u_{j})\).

Here,

  • \(\qhat_j\in\RR^{r}\) is the \(j\)-th iteration of the model state, and

  • \(\u_j\in\RR^{m}\) is the (optional) corresponding input.

The structure of \(\fhat\) is specified through the operators attribute.

Parameters
operatorslist of opinf.operators objects

Operators comprising the terms of the model.

Properties

A_

opinf.operators.LinearOperator (or None).

B_

opinf.operators.InputOperator (or None).

G_

opinf.operators.CubicOperator (or None).

H_

opinf.operators.QuadraticOperator (or None).

N_

opinf.operators.StateInputOperator (or None).

c_

opinf.operators.ConstantOperator (or None).

data_matrix_

\(k \times d(r, m)\) data matrix, e.g., \(\D = [~ \mathbf{1}~~ \widehat{\Q}\trp~~ (\widehat{\Q}\odot\widehat{\Q})\trp~~ \U\trp~]\).

input_dimension

Dimension \(m\) of the input (zero if there are no inputs).

operator_matrix_

\(r \times d(r, m)\) operator matrix, e.g., \(\Ohat = [~\chat~~\Ahat~~\Hhat~~\Bhat~]\).

operator_matrix_dimension

Number of columns \(d(r, m)\) of the operator matrix \(\Ohat\) and the data matrix \(\D\), i.e., the number of unknowns in the Operator Inference regression problem for each system mode.

operators

Operators comprising the terms of the model.

state_dimension

Dimension \(r\) of the state.

Methods

copy

Make a copy of the model.

fit

Learn the model operators from data.

galerkin

Construct a reduced-order model by taking the (Petrov-)Galerkin projection of each model operator.

jacobian

Sum the state Jacobian of each model operator.

load

Load a serialized model from an HDF5 file, created previously from the save() method.

predict

Step forward the discrete dynamical system niters steps.

rhs

Evaluate the right-hand side of the model by applying each operator and summing the results.

save

Serialize the model, saving it in HDF5 format.

stack_trajectories

Translate a collection of state trajectories and (optionally) inputs to arrays that are appropriate arguments for fit().