FiniteDifferenceMethod Enumeration |
When numerically computing the derivative of a function, this will indicate if the
finite difference method should be forward, backwards or central.
Namespace:
AGI.Foundation.NumericalMethods
Assembly:
AGI.Foundation.Core (in AGI.Foundation.Core.dll) Version: 24.2.419.0 (24.2.419.0)
Syntax public enum FiniteDifferenceMethod
Public Enumeration FiniteDifferenceMethod
public enum class FiniteDifferenceMethod
type FiniteDifferenceMethod
Members
| Member name | Description |
---|
| Forward |
A forward difference method will be performed. When perturbing a variable, the perturbation
will be added to the current value of that variable. The total number of function evaluations
that will be performed to solve for the Jacobian of a multivariable function will be 1 plus
the number of variables.
|
| Backward |
A backwards difference method will be performed. When perturbing a variable, the perturbation
will be subtracted to the current value of that variable. The total number of function evaluations
that will be performed to solve for the Jacobian of a multivariable function will be 1 plus
the number of variables.
|
| Signed |
The difference will be the signed perturbation value. The total number of function evaluations
that will be performed to solve for the Jacobian of a multivariable function will be 1 plus
the number of variables.
|
| Central |
A central difference method will be performed. Two function evaluations will be performed
for each variable to determine the change of the function values. One run will be done
where half of the perturbation value is subtracted from the nominal value of the variable, and a
second run where half of the perturbation value is added to the variable. The total number of function evaluations
that will be performed to solve for the Jacobian of a multivariable function will be 1 plus
twice the number of variables.
|
See Also