lines 6-60 of file: xrst/theory/taylor_ode.xrst

{xrst_begin taylor_ode}

AD Theory for Solving ODE's Using Taylor's Method
#################################################

Problem
*******
We are given an initial value problem for
:math:`y : \B{R} \rightarrow \B{R}^n`; i.e., we know
:math:`y(0) \in \B{R}^n`
and we know a function :math:`g : \B{R}^n \rightarrow \B{R}^n`
such that :math:`y^1 (t) = g[ y(t) ]`
where :math:`y^k (t)` is the *k*-th derivative of :math:`y(t)`.

z(t)
****
We define the function
:math:`z : \B{R} \rightarrow \B{R}^n` by
:math:`z(t) = g[ y(t) ]`.
Given the Taylor coefficients
:math:`y^{(k)} (t)` for :math:`k = 0 , \ldots , p`, we can
compute :math:`z^{(p)} (t)` using forward mode AD on the function
:math:`g(y)`; see :ref:`forward_order-name` .
It follows from :math:`y^1 (t) = z(t)` that
:math:`y^{p+1} (t) = z^p (t)`

.. math::

   y^{(p+1)} (t) = z^{(p)} (t) / (k + 1)

where :math:`y^{(k)} (t)` is the *k*-th order Taylor coefficient
for :math:`y(t)`; i.e., :math:`y^k (t) / k !`.
Starting with the known value :math:`y^{(0)} (t)`,
this gives a prescription for computing :math:`y^{(k)} (t)` for
any :math:`k`.

Taylor's Method
***************
The *p*-th order Taylor method for approximates

.. math::

   y( t + \Delta t )
   \approx
   y^{(0)} (t) + y^{(1)} (t) \Delta t + \cdots + y^{(p)} (t)  \Delta t^p

{xrst_toc_hidden
   example/general/taylor_ode.cpp
}
Example
*******
The file :ref:`taylor_ode.cpp-name` contains an example and test
of this method.

{xrst_end taylor_ode}