view scripts/ode/private/runge_kutta_45_dorpri.m @ 27918:b442ec6dda5c

use centralized file for copyright info for individual contributors * COPYRIGHT.md: New file. * In most other files, use "Copyright (C) YYYY-YYYY The Octave Project Developers" instead of tracking individual names in separate source files. The motivation is to reduce the effort required to update the notices each year. Until now, the Octave source files contained copyright notices that list individual contributors. I adopted these file-scope copyright notices because that is what everyone was doing 30 years ago in the days before distributed version control systems. But now, with many contributors and modern version control systems, having these file-scope copyright notices causes trouble when we update copyright years or refactor code. Over time, the file-scope copyright notices may become outdated as new contributions are made or code is moved from one file to another. Sometimes people contribute significant patches but do not add a line claiming copyright. Other times, people add a copyright notice for their contribution but then a later refactoring moves part or all of their contribution to another file and the notice is not moved with the code. As a practical matter, moving such notices is difficult -- determining what parts are due to a particular contributor requires a time-consuming search through the project history. Even managing the yearly update of copyright years is problematic. We have some contributors who are no longer living. Should we update the copyright dates for their contributions when we release new versions? Probably not, but we do still want to claim copyright for the project as a whole. To minimize the difficulty of maintaining the copyright notices, I would like to change Octave's sources to use what is described here: https://softwarefreedom.org/resources/2012/ManagingCopyrightInformation.html in the section "Maintaining centralized copyright notices": The centralized notice approach consolidates all copyright notices in a single location, usually a top-level file. This file should contain all of the copyright notices provided project contributors, unless the contribution was clearly insignificant. It may also credit -- without a copyright notice -- anyone who helped with the project but did not contribute code or other copyrighted material. This approach captures less information about contributions within individual files, recognizing that the DVCS is better equipped to record those details. As we mentioned before, it does have one disadvantage as compared to the file-scope approach: if a single file is separated from the distribution, the recipient won't see the contributors' copyright notices. But this can be easily remedied by including a single copyright notice in each file's header, pointing to the top-level file: Copyright YYYY-YYYY The Octave Project Developers See the COPYRIGHT file at the top-level directory of this distribution or at https://octave.org/COPYRIGHT.html. followed by the usual GPL copyright statement. For more background, see the discussion here: https://lists.gnu.org/archive/html/octave-maintainers/2020-01/msg00009.html Most files in the following directories have been skipped intentinally in this changeset: doc libgui/qterminal liboctave/external m4
author John W. Eaton <jwe@octave.org>
date Mon, 06 Jan 2020 15:38:17 -0500
parents 00f796120a6d
children 1891570abac8
line wrap: on
line source

## Copyright (C) 2013-2019 The Octave Project Developers
##
## See the file COPYRIGHT.md in the top-level directory of this distribution
## or <https://octave.org/COPYRIGHT.html/>.
##
##
## This file is part of Octave.
##
## Octave is free software: you can redistribute it and/or modify it
## under the terms of the GNU General Public License as published by
## the Free Software Foundation, either version 3 of the License, or
## (at your option) any later version.
##
## Octave is distributed in the hope that it will be useful, but
## WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
## GNU General Public License for more details.
##
## You should have received a copy of the GNU General Public License
## along with Octave; see the file COPYING.  If not, see
## <https://www.gnu.org/licenses/>.

## -*- texinfo -*-
## @deftypefn  {} {[@var{t_next}, @var{x_next}] =} runge_kutta_45_dorpri (@var{@@fun}, @var{t}, @var{x}, @var{dt})
## @deftypefnx {} {[@var{t_next}, @var{x_next}] =} runge_kutta_45_dorpri (@var{@@fun}, @var{t}, @var{x}, @var{dt}, @var{options})
## @deftypefnx {} {[@var{t_next}, @var{x_next}] =} runge_kutta_45_dorpri (@var{@@fun}, @var{t}, @var{x}, @var{dt}, @var{options}, @var{k_vals})
## @deftypefnx {} {[@var{t_next}, @var{x_next}] =} runge_kutta_45_dorpri (@var{@@fun}, @var{t}, @var{x}, @var{dt}, @var{options}, @var{k_vals}, @var{t_next})
## @deftypefnx {} {[@var{t_next}, @var{x_next}, @var{x_est}] =} runge_kutta_45_dorpri (@dots{})
## @deftypefnx {} {[@var{t_next}, @var{x_next}, @var{x_est}, @var{k_vals_out}] =} runge_kutta_45_dorpri (@dots{})
##
## This function can be used to integrate a system of ODEs with a given initial
## condition @var{x} from @var{t} to @var{t+dt} with the Dormand-Prince method.
## For the definition of this method see
## @url{http://en.wikipedia.org/wiki/Dormand%E2%80%93Prince_method}.
##
## First input argument is the function describing the system of ODEs to be
## integrated.
##
## Second input parameter is the first extreme of integration interval.
##
## Third input argument is the initial condition of the system.
##
## Fourth input argument is the timestep, that is the length of the
## integration interval.
##
## Fifth input parameter is optional and describes a set of options useful to
## adapt the computation to what is needed.
##
## Sixth input parameter is optional and describes the Runge-Kutta evaluations
## of the previous step to use in an FSAL scheme.
##
## Seventh input parameter is optional and is the time (@var{t_next}) to
## integrate to.  The default is @code{@var{t} + @var{dt}}.
##
## First output argument is the final integration time value.
##
## Second output parameter is the higher order computed solution at time
## @var{t_next} (local extrapolation).
##
## Third output parameter is a lower order solution for the estimation of the
## error.
##
## Fourth output parameter is matrix containing the Runge-Kutta evaluations
## to use in an FSAL scheme or for dense output.
## @end deftypefn

function [t_next, x_next, x_est, k] = runge_kutta_45_dorpri (fun, t, x, dt,
                                                             options = [],
                                                             k_vals = [],
                                                             t_next = t + dt)

  ## Reference: Hairer, Ernst; Nørsett, Syvert Paul; Wanner, Gerhard (2008),
  ## Solving ordinary differential equations I: Nonstiff problems,
  ## Berlin, New York: Springer-Verlag, ISBN 978-3-540-56670-0
  persistent a = [0           0          0           0        0          0;
                  1/5         0          0           0        0          0;
                  3/40        9/40       0           0        0          0;
                  44/45      -56/15      32/9        0        0          0;
                  19372/6561 -25360/2187 64448/6561 -212/729  0          0;
                  9017/3168  -355/33     46732/5247  49/176  -5103/18656 0];
  persistent b = [0, 1/5, 3/10, 4/5, 8/9, 1, 1];
  persistent c = [35/384, 0, 500/1113, 125/192, -2187/6784, 11/84];
  persistent c_prime = [5179/57600, 0, 7571/16695, 393/640, ...
                        -92097/339200, 187/2100, 1/40];

  s = t + dt * b;
  cc = dt * c;
  aa = dt * a;
  k = zeros (rows (x), 7);

  if (! isempty (options))   # extra arguments for function evaluator
    args = options.funarguments;
  else
    args = {};
  endif

  if (! isempty (k_vals))    # k values from previous step are passed
    k(:,1) = k_vals(:,end);  # FSAL property
  else
    k(:,1) = feval (fun, t, x, args{:});
  endif

  k(:,2) = feval (fun, s(2), x + k(:,1)   * aa(2, 1).'  , args{:});
  k(:,3) = feval (fun, s(3), x + k(:,1:2) * aa(3, 1:2).', args{:});
  k(:,4) = feval (fun, s(4), x + k(:,1:3) * aa(4, 1:3).', args{:});
  k(:,5) = feval (fun, s(5), x + k(:,1:4) * aa(5, 1:4).', args{:});
  k(:,6) = feval (fun, s(6), x + k(:,1:5) * aa(6, 1:5).', args{:});

  ## compute new time and new values for the unknowns
  ## t_next = t + dt;
  x_next = x + k(:,1:6) * cc(:);  # 5th order approximation

  ## if the estimation of the error is required
  if (nargout >= 3)
    ## new solution to be compared with the previous one
    k(:,7) = feval (fun, t_next, x_next, args{:});
    cc_prime = dt * c_prime;
    x_est = x + k * cc_prime(:);
  endif

endfunction