odeset

Create or update ODE options.

📝Syntax

options = odeset()
options = odeset(name, value)
options = odeset(oldOptions, name, value)

📄Description

odeset creates an options structure for ODE solvers.

Group	Options	Purpose
Tolerances	RelTol, AbsTol, NormControl	Control the adaptive error test.
Steps and output	InitialStep, MaxStep, MinStep, Refine, OutputFcn, OutputSel, Stats	Control public output points, output callbacks, and statistics.
Events and constraints	Events, NonNegative	Locate zeros and constrain selected components.
Stiff systems	Mass, Jacobian, JPattern, JConstant, BDF, MaxOrder, MassSingular, MStateDependence, MvPattern	Provide structural information to stiff or implicit solvers.

Common options include RelTol, AbsTol, InitialStep, MaxStep, MinStep, Refine, Stats, Events, OutputFcn, OutputSel, NonNegative, Mass, Jacobian, JPattern, JConstant, BDF, MaxOrder, MassSingular, MStateDependence, and MvPattern. Explicit step-size options must be positive.

When the time span has only two values, Refine inserts additional output points inside each accepted step. With longer time spans, requested time points define the public output points. OutputFcn is called at these refined or requested output points; returning scalar true stops integration at the current output point. Stats set to on prints step and evaluation counters. NormControl set to on uses a vector norm in the adaptive error test. Vectorized set to on allows vectorized right-hand side calls during finite-difference Jacobian construction. JPattern supplies the nonzero pattern used by stiff entries and the implicit entry when a finite-difference Jacobian is needed. JConstant set to on reuses the Jacobian inside each linearly implicit step.

💡Examples

Refined output points.

options = odeset('Refine', 4, 'MaxStep', 0.5);
[t, y] = ode45(@(t,y) y, [0 1], 1, options)

Stats and a Jacobian for a stiff entry point.

f = @(t,y) -1000 * (y - cos(t)) - sin(t);
options = odeset('Jacobian', -1000, 'Stats', 'on');
[t, y] = ode15s(f, [0 0.5], 1, options)

Vector norm control.

options = odeset('NormControl', 'on');
[t, y] = ode45(@(t,y) [y(1); -2*y(2)], [0 1], [1; 1], options)

Finite-difference Jacobian pattern.

pattern = [1 0; 0 1];
options = odeset('JPattern', pattern, 'Vectorized', 'on');
[t, y] = ode15s(@(t,y) [-10*y(1,:); -20*y(2,:)], [0 0.2], [1; 2], options)

Constant Jacobian hint.

options = odeset('Jacobian', -25, 'JConstant', 'on');
[t, y] = ode15s(@(t,y) -25*y, [0 0.2], 1, options)

Output function at requested points.

out = @(t,y,flag) false;
options = odeset('OutputFcn', out);
[t, y] = ode45(@(t,y) y, [0 0.25 0.5], 1, options)

🔗See Also

odeget ode45 ode15s

🕔Version History

Version	Description
2.0.0	initial version

Edit this page on GitHub