Prev

Next

Index-> contents reference index search external

Up-> pycppad ad_function forward

pycppad-> install get_started.py example ad_variable ad_function two_levels.py runge_kutta_4 whats_new license

ad_function-> independent adfun abort_recording forward reverse jacobian hessian optimize

forward-> forward_0.py forward_1.py

Headings-> Syntax Purpose x_k y_k f p x_p y_p Example

@(@\newcommand{\B}[1]{{\bf #1}} \newcommand{\R}[1]{{\rm #1}}@)@ Forward Mode: Derivative in One Domain Direction
Syntax
y_p = f.forward(p, x_p)

Purpose
We use @(@ F : \B{R}^n \rightarrow \B{R}^m @)@ to denote the function corresponding to the adfun object f . Given the p-th order Taylor expansion for a function @(@ X : \B{R} \rightarrow \B{R}^n @)@, this function can be used to compute the p-th order Taylor expansion for the function @(@ Y : \B{R} \rightarrow \B{R}^m @)@ defined by @[@ Y(t) = F [ X(t) ] @]@

x_k
For @(@ k = 0 , \ldots , p @)@, we use @(@ x^{(k)} @)@ to denote the value of x_k in the most recent call to
     f.forward(k, x_k)
including @(@ x^{(p)} @)@ as the value x_p in this call. We define the function @(@ X(t) @)@ by @[@ X(t) = x^{(0)} + x^{(1)} * t + \cdots + x^{(p)} * t^p @]@

y_k
For @(@ k = 0 , \ldots , p @)@, we use @(@ y^{(k)} @)@ to denote the Taylor coefficients for @(@ Y(t) = F[ X(t) ] @)@ expanded about zero; i.e., @[@ \begin{array}{rcl} y^{(k)} & = & Y^{(k)} (0) / k ! \\ Y(t) & = & y^{(0)} + y^{(1)} * t + \cdots + y^{(p)} * t^p + o( t^p ) \end{array} @]@ where @(@ o( t^p ) / t^p \rightarrow 0 @)@ as @(@ t \rightarrow 0 @)@. The coefficient @(@ y^{(p)} @)@ is equal to the value y_p returned by this call.

f
The object f must be an adfun object. We use level for the AD ad level of this object.

p
The argument p is a non-negative int. It specifies the order of the Taylor coefficient for @(@ Y(t) @)@ that is computed.

x_p
The argument x_p is a numpy.array with one dimension (i.e., a vector) with length equal to the domain size n for the function f . It specifies the p-th order Taylor coefficient for @(@ X(t) @)@. If the AD level for f is zero, all the elements of x_p must be either int or instances of float. If the AD level for f is one, all the elements of x_p must be a_float objects.

y_p
The return value y_p is a numpy.array with one dimension (i.e., a vector) with length equal to the range size m for the function f . It is set to the p-th order Taylor coefficient for @(@ Y(t) @)@. If the AD level for f is zero, all the elements of y_p will be instances of float. If the AD level for f is one, all the elements of y_p will be a_float objects.

Example

forward_0.py	Forward Order Zero: Example and Test
forward_1.py	Forward Order One: Example and Test

---

Input File: pycppad/adfun.cpp