options = nelson.ode.options.ODE23tb()
options = nelson.ode.options.ODE23tb(name, value)
nelson.ode.options.ODE23tb creates a compatible option class for the 'ode23tb' solver value used by the ode object workflow.
| Option group | Names | Purpose |
|---|---|---|
| Steps | InitialStep, MaxStep, MinStep | Bound the adaptive step size selection. |
| Error control | NormControl | Switch between componentwise and norm based error control. |
| Evaluation | Vectorization | Declare that the ODE function accepts matrices of states. |
| Output | OutputFcn, OutputSelection | Select output callbacks and returned components. |
The 'ode23tb' solver value uses an implicit Runge-Kutta method combining a trapezoidal rule stage with a backward differentiation stage, effective for stiff problems at crude tolerances.
Supported properties are InitialStep, MaxStep, MinStep, NormControl, OutputFcn, OutputSelection, and Vectorization. InitialStep, MaxStep, and MinStep are positive scalars bounding the adaptive step size; their default value is empty, which lets the solver choose them automatically. NormControl accepts 'on' or 'off' (default 'off') and enables error control based on the norm of the solution instead of componentwise control. Vectorization accepts 'on' or 'off' (default 'off') and declares that the ODE function can evaluate several columns of states at once. OutputFcn is a function handle called on each output point (default empty). OutputSelection is a vector of indices selecting which solution components are passed to the output function (default empty, all components). The default Refine value for this solver is 1.
options = nelson.ode.options.ODE23tb('Vectorization', 'on');
problem = ode('ODEFcn', @(t,y) -100 * y, 'InitialValue', 1, ...
'SolverOptions', options);
result = solve(problem, 0, 1)| Version | Description |
|---|---|
| 2.0.0 | initial version |