function [hessf, err] = tapas_riddershessian(f, x, varargin)
% Calculates the hessian of the function f at point x according to Ridders' method:
%
% Ridders, CJF. (1982). Accurate computation of F'(x) and F'(x) F''(x). Advances in Engineering
% Software, 4(2), 75-6.
%
% INPUT:
% f Function handle of a real function of n real variables which are passed as
% *one* vector with n elements
% x Point at which to differentiate f
%
% OUTPUT:
% hessf Hessian of f at x
% err Error estimates
%
% OPTIONS:
% Optionally, the third argument of the function can be a structure containing further
% settings for Ridder's method.
%
% varargin{1}.init_h Initial finite difference (default: 1)
% varargin{1}.div Divisor used to reduce h on each step (default: 1.2)
% varargin{1}.min_steps Minimum number of steps in h (default: 3)
% varargin{1}.max_steps Maximum number of steps in h (default: 100)
% varargin{1}.tf Terminate if last step worse than preceding by a factor of tf
% (default: 2)
%
% --------------------------------------------------------------------------------------------------
% Copyright (C) 2012-2013 Christoph Mathys, TNU, UZH & ETHZ
%
% This file is released under the terms of the GNU General Public Licence (GPL), version 3. You can
% redistribute it and/or modify it under the terms of the GPL (either version 3 or, at your option,
% any later version). For further details, see the file COPYING or <http://www.gnu.org/licenses/>.
n = length(x);
% Defaults
options.init_h = 1;
options.div = 1.2;
options.min_steps = 3;
options.max_steps = 100;
options.tf = 2;
hessf = NaN(n);
err = NaN(n);
% Overrides
if nargin > 2
options_ovr = varargin{1};
if isfield(options_ovr,'init_h')
options.init_h = options_ovr.init_h;
end
if isfield(options_ovr,'div')
options.div = options_ovr.div;
end
if isfield(options_ovr,'min_steps')
options.min_steps = options_ovr.min_steps;
end
if isfield(options_ovr,'max_steps')
options.max_steps = options_ovr.max_steps;
end
if isfield(options_ovr,'tf')
options.tf = options_ovr.tf;
end
end
% Check if f and x match
try
f(x);
catch err
error('tapas:hgf:ridders:CannotEvalFun', 'Function cannot be evaluated at differentiation point.');
end
% First: diagonal elements
% Loop through argument variables
for i = 1:n
% Construct filehandle to be passed to riddersdiff2
fxi = @(xi) fxi(f,x,i,xi);
% Calculate derivative
[hessf(i,i), err(i,i)] = tapas_riddersdiff2(fxi,x(i),options);
end
% Second: off-diagonal elements
% Loop through argument variables
for i = 2:n % rows
for j = 1:i-1 % columns
% Construct filehandle to be passed to riddersdiffcross
fxixj = @(xixj) fxixj(f,x,i,j,xixj);
% Calculate cross-derivative
[hessf(i,j), err(i,j)] = tapas_riddersdiffcross(fxixj,[x(i),x(j)],options);
hessf(j,i) = hessf(i,j);
err(j,i) = err(i,j);
end
end
end
function fxi = fxi(f,x,i,xi)
xx = x;
xx(i) = xi;
fxi = f(xx);
end
function fxixj = fxixj(f,x,i,j,xixj)
xx = x;
xx(i) = xixj(1);
xx(j) = xixj(2);
fxixj = f(xx);
end