Change language to:
English - 日本語 - Português - Русский

See the recommended documentation of this function

Aide de Scilab >> Calcul différentiel, Intégration > daeoptions

# daeoptions

set options for dae solver

### Calling Sequence

`daeoptions()`

### Description

Sets a variable `%DAEOPTIONS` that is used in function `dae` to set solver options.

This daeoptions function interactively displays a window which sets various options of the dae solver.

 The `dae` function checks if this variable exists and in this case it uses it. For using default values you should clear this variable. Note that `daeoptions` does not create this variable. To create it you must execute the command line displayed by `daeoptions`.

The variable `%DAEOPTIONS` is a list with the following elements:

`list(tstop,imode,band,maxstep,stepin,nonneg,isest)`

The default value is:

`list([],0,[],[],[],0,0)`

The meaning of the elements is described below.

solver

if it is 0, `dae` will use the dassl/dasrt solver, and if it is 1, `dae` will use daskr.

tstop

a real scalar or an empty matrix, gives the maximum time for which `g` is allowed to be evaluated. An empty matrix means "no limits" imposed for time.

imode

if it is 0, `dae` returns only the user specified time point values, and if it is 1, `dae` returns its intermediate computed values.

band

a two components vector which give the definition `[ml,mu]` of band matrix computed by `jac`;

`r(i - j + ml + mu + 1,j)` = `dg(i)/dy(j)+cj*dg(i)/dydot(j)`. If `jac` returns a full matrix set `band=[]`

maxstep

a scalar or an empty matrix, the maximum step size, empty matrix means "no limitation".

stepin

a scalar or an empty matrix, the minimum step size, empty matrix means "not specified".

nonneg

a scalar, must be set to 0 if the solution is known to be non negative . In the other case it must be set to 1.

isest

a scalar, must be set to 0 if the given initial condition is compatible: `g(t0,x0,xdot0)=0`. If `xdot0` is just an estimation, then set to 1 if you are using dassl, or set to `[+-1, ..., +-1]`, with: `1` if y(i) is a differential variable and `-1` if y(i) is an algebraic variable (if its derivatives do not appear explicitly in the function g(t, y, ydot)).

method

a scalar, direct / Krylov method. Set to `1` and provide a routine `psol` in `dae()` if you want the solver to use Krylov iterations, else (daskr's direct method) set to `0`.

Kry_params

a vector, treat as dummy argument if you have set method=0. Otherwise, set to `[]` or to `[maxl kmp nrmax epli]`, where: - `maxl` = maximum number of iterations in the GMRes algorithm (default min(5, neq)), - `kmp` = number of vectors on which orthogonalization is done in the GMRes algorithm (default maxl), - `nrmax` = maximum number of restarts of the GMRes algorithm per nonlinear iteration (default 5), - `epli` = convergence test constant in GMRes algorithm (default 0.05).

init

a scalar, treat as dummy argument if you have set isest=0. Set to 1 if the solver should stop right after computation of the initial conditions, else set to 0.

precond

a scalar, preconditioner computation and LU-factorization routine for `psol`. Treat as dummy argument if method=0. Set to 1 and provide a `pjac` routine in `dae` if the external `psol` should use a specific routine, else set to 0.

control

a vector, treat as dummy argument if you have set isest=0if you wish to control errors locally on all the variables then set to []. Otherwise, set to [+-1, ..., +-1], 1 if `y(i)` is a differential variable and -1 if `y(i)` is an algebraic variable (if its derivatives do not appear explicitly in the function `g(t, y, ydot)`).

heuristic

a vector, heuristic parameters. Treat as dummy argument if isest=0. Otherwise, set to [] or to `[mxnit mxnj mxnh lsoff stptol epinit]`, where: - `mxnit` = maximum number of Newton iterations per Jacobian or preconditioner evaluation (default 5 if info(8)=0, 15 otherwise), - `mxnj` = maximum number of Jacobian or preconditioner evaluations (default 6 if info(8)=0, 2 otherwise), - `mxnh` = maximum number of values of the artificial stepsize parameter h to be tried if info(7) ≠ [] (default 5), - `lsoff` = flag to turn off the linesearch algorithm (lsoff = 0 means linesearch is on, lsoff = 1 means it is turned off) (default 0), - `stptol` = minimum scaled step in linesearch algorithm (default (unit roundoff)^(2/3)), - `epinit` = swing factor in the Newton iteration convergence test (default 0.01).

verbosity

a scalar, verbosity. Set to 1 for minimal extra printing, 2 for full printing, else set to 0.