steep_desc2.m

BASIC MULTIDIMENSIONAL GRADIENT METHODS

% Program: steep_desc2.m
% Title: Steepest-descent algorithm without line search
% Description: Implements the steepest-descent method
% described in Algorithm 5.2.  Parameter alpha is determined 
% using the closed-form formula in Eq. (5.7) instead of
% a line search.
% Theory: See Practical Optimization Sec. 5.2.4.
% Input: 
%   fname: objective function
%   gname: gradient of the objective function
%      x0: initial point
%    epsi: termination tolerance
% Output:   
%      xs: solution point
%      fs: objective function evaluated at xs.
%       k: number of iterations at convergence
% Example:
% Find the minimum of the Himmelblau function
%    f = (x1^2 + x2 - 11)^2 + (x1 + x2^2 - 7)^2
% with initioal point x0 = [6 6]' and termination tolerance 
% epsi = 1e-6.
% Solution:
% Execute the command
%   [xs,fs,k] = steep_desc2('f_himm','g_himm',[6 6]',1e-6)
% Notes:
% 1. The program can be applied to any customized function
%    by defining the function of interest and its gradient 
%    in m-files.
% ========================================================
function [xs,fs,k] = steep_desc2(fname,gname,x0,epsi)
disp(' ')
disp('Program steep_desc2.m')
k = 1;
xk = x0;
fk = feval(fname,xk);
gk = feval(gname,xk);
ah = 1;
fh = feval(fname,xk-ah*gk);
gk2 = gk'*gk;
ak = (gk2*ah^2)/(2*(fh-fk+ah*gk2));
adk = -ak*gk;
er = norm(adk);
while er >= epsi,
   xk = xk + adk;
   fk = feval(fname,xk);
   gk = feval(gname,xk);
   ah = ak;
   fh = feval(fname,xk-ah*gk);
   gk2 = gk'*gk;
   ak = (gk2*ah^2)/(2*(fh-fk+ah*gk2));
   adk = -ak*gk;
   er = norm(adk);
   k = k + 1;
end
format long
disp('Solution point:')
xs = xk + adk
disp('Objective function at the solution point:')
fs = feval(fname,xs)
format short
disp('Nnumber of iterations performed:')
k