Main Content

Solver Jacobian method

Method implicit solvers use to compute solver Jacobian

Model Configuration Pane: Solver

Description

The Solver Jacobian method parameter specifies how implicit solvers compute the solver Jacobian. By default, the software chooses the solver Jacobian method. For most models, the software chooses an appropriate method.

For details, see Choose Jacobian Method for Implicit Solvers.

For information about the available implicit solvers, see Variable Step Solvers in Simulink.

Dependencies

To enable this parameter, configure your model with one of these parameter combinations:

  • Set the solver Type to Variable-step and use the Solver parameter to select one of these implicit solvers:

    • ode15s (stiff/NDF)

    • ode23s (stiff/Mod. Rosenbrock)

    • ode23t (mod. stiff/Trapezoidal)

    • ode23tb (stiff/TR-BDF2)

    • daessc (DAE solver for Simscape™)

  • Set the solver Type to Variable-step. For the Solver parameter, select odeN (Nonadaptive). Then, for the Integration method parameter, select ode14x (extrapolation) or ode1be (Backward Euler).

  • Set the solver Type to Fixed-step and set the Solver parameter to ode14x (extrapolation) or ode1be (Backward Euler).

Settings

auto (default) | Sparse perturbation | Full perturbation | Sparse analytical | Full analytical
auto

The software chooses the solver Jacobian method by analyzing the model and the selected solver. In most cases, the software chooses an appropriate method.

Sparse perturbation

The solver changes state values to compute the partial state derivatives for the solver Jacobian. The sparse perturbation method uses the solver Jacobian sparsity pattern to reduce the number of calculations.

Data store blocks, such as the Data Store Memory block do not support sparse perturbation. If you select this option in a model that contains one or more data store blocks, the software uses the full perturbation method.

Full perturbation

The solver changes state values to compute the partial state derivatives for the solver Jacobian. The full perturbation method solves the full set of perturbation equations for the system.

In model hierarchies that contain one or more models configured to use a local solver, the parameter must use this option or auto if the top model or any referenced model uses an implicit solver.

Sparse analytical

The solver computes the solver Jacobian by solving analytical equations and uses the sparsity pattern to reduce the number of calculations.

This option is not supported for rapid accelerator simulations. If you select this option, the software uses the full perturbation method in rapid accelerator simulations.

Full analytical

The solver computes the solver Jacobian by solving analytical equations.

For details, see Choose Jacobian Method for Implicit Solvers.

Programmatic Use

Parameter: SolverJacobianMethodControl
Type: string | character vector
Value: "auto" | "SparsePerturbation" | "FullPerturbation" | "SparseAnalytical" | "FullAnalytical"
Default: 'auto'

Version History

Introduced in R2010b