home *** CD-ROM | disk | FTP | other *** search
Text File | 1994-06-05 | 3.6 KB | 146 lines | [MATS/MATL] |
- echo off;
- % NUMERICAL METHODS: MATLAB Programs, (c) John H. Mathews 1994
- % To accompany the text:
- % NUMERICAL METHODS for Mathematics, Science and Engineering, 2nd Ed, 1992
- % Prentice Hall, Englewood Cliffs, New Jersey, 07632, U.S.A.
- % This free software is complements of the author.
-
- % Algorithm 8.4, (Steepest Descent or Gradient Method).
- % Section 8.1, Minimization of a Function, Page 418
- echo on; clc; format short; hold off; clear
-
- % This program finds the minimum of a function Fn(X),
-
- % where X is an n-dimensional vector.
-
- % The technique is the gradient method for n-dimensions.
-
- % The function Fn(X) is to be placed in two M-files.
-
- % The function f.m is for graphing and Fn.m is for computing.
-
- % The gradient of Fn(X) is to be placed in the M-file Gn.m
-
- pause % Press any key to continue.
-
- clc;
- % function z = f(x,y)
- % z = x.^2 - 4.*x + y.^2 - y - x.*y;
-
- % function z = Fn(X)
- % x = X(1); y = X(2);
- % z = x^2 - 4*x + y^2 - y - x*y;
-
- % function Z = Gn(X)
- % x = X(1); y = X(2);
- % G = [2*x-4-y, 2*y-1-x];
- % if norm(G)~=0,
- % Z = -G/norm(G);
- % else
- % Z = -G;
- % end;
-
- delete f.m
- diary f.m; disp('function z = f(x,y)');...
- disp('z = x.^2 - 4.*x + y.^2 - y - x.*y;');...
- diary off;
-
- delete Fn.m
- diary Fn.m; disp('function z = Fn(X)');...
- disp('x = X(1); y = X(2);');...
- disp('z = x^2 - 4*x + y^2 - y - x*y;');...
- diary off;
-
- delete Gn.m
- diary Gn.m; disp('function Z = Gn(X)');...
- disp('x = X(1); y = X(2);');...
- disp('G = [2*x-4-y, 2*y-1-x];');...
- disp('if norm(G)~=0,');...
- disp(' Z = -G/norm(G);');...
- disp('else');...
- disp(' Z = -G;');...
- disp('end;');...
- diary off;
-
- % Remark. f.m Fn.m Gn.m grads.m are used for Algorithm 8.4
- % Remark. f(x,y) is used only to graph 2-dimensional functions.
- f(0,0); Fn([0 0]); Gn([0 0]);
- pause % Press any key to continue.
-
- clc;
- % The example is 2-dimensional, hence a surface z = f(x,y)
-
- % can be graphed. (Omit this section for higher dimensions.)
-
- % The endpoints of the rectangle [a,b] x [c,d] are a,b,c,d.
-
- % Place the maximum number of iterations in max1
-
- % Place the tolerances in delta and epsilon.
-
- a = 0;
- b = 4;
- c = 0;
- d = 4;
- max1 = 50;
- delta = 1e-5;
- epsilon = 1e-7;
-
- pause % Press any key to enter the starting vertices.
-
- clc; clg;
- hs = (b-a)/16;
- ks = (d-c)/16;
- [Xs,Ys] = meshdom(a:hs:b, c:ks:d);
- Zs = f(Xs,Ys); % User defined function.
- mesh(Zs);...
- title('To search for a minimum of z = f(x,y).');...
- shg; pause % Press any key to continue.
-
- clc;
-
- % The gradient method is used to find the minimum
-
- % of the n-dim function Fn(X). for convenience,
-
- % For convenience, Fn(X) can be expressed using
-
- % X = (x1,x2,x3,x4,x5,x6) and the variables
-
- % x = x1 , y = x2 , z = x3 , u = x4 , v = x5 , w = x6
-
- % You must supply the starting point P0(1:n)
-
- P0 = [0.0 0.0];
-
- pause % Press any key to find the minimum of Fn(X).
-
- clc; clg;
- a0 = 0;
- b0 = 3.5;
- c0 = 0;
- d0 = 2.5;
- hold off;
- axis([a0 b0 c0 d0]);...
- plot(P0(1),P0(2),'o');...
- hold on;...
- plot([a0 b0],[0 0],'b',[0 0],[c0 d0],'b');...
- xlabel('x');...
- ylabel('y');...
- title('Iteration points for the gradient method.');...
- grid;...
- shg; pause % Press any key to continue.
-
- [P0,y0,h,err,P] = grads('Fn','Gn',P0,max1,delta,epsilon,1);
-
- format long;
- Mx1 = 'The coordinates of the point P are:';
- Mx2 = 'Error bound for the coordinates of P:';
- Mx3 = 'The minimum value F(P) is:';
- Mx4 = 'Error bound for F(P) is:';
- clc,echo off,diary output,...
- disp(''),disp(Mx1),disp(P0),disp(Mx2),...
- disp(['± ',num2str(h),' ± ',num2str(h)]),disp(''),...
- disp(Mx3),disp(y0),disp(Mx4),disp(['± ',num2str(err)]),...
- diary off,echo on
-