Why my code is not working?

1 view (last 30 days)
Biza Ferreira on 9 Jan 2015
Commented: Jan on 9 Jan 2015
I have a small problem, I need resover a problem through the quadratic penalty method, I have done the objective function and constraints of equality and inequality, but now I have a problem in the file you use the fminsearch function, does anyone can help me. I did the following code.
func.m
if true
function [y,rd,ri]=func(x)
A=[5.22 3.11 0.01 1.02 0.00;0.00 2.12 1.03 0.05 1.04;1.12 0.19 3.23 1.05 2.03;2.89 3.76 0.74 4.32 0.00;0.01 1.99 3.88 2.01 5.55];
y=x'*A*x;
%constraints inequality (rd>=0)
rd=[x(1);-x(1)+1;x(2);-x(2)+1;x(3);-x(3)+1;x(4);-x(4)+1;x(5);-x(5)+1];
%constraints (ri=0)
ri=[sum(x)-5];
end
end
funobj_qpm.m
if true
function y=funobj_qpm(funobj,miu,x)
[yy,rd,ri]=funobj(x);
y=yy+(1/(2*miu))*sum(ri.^2)+(1/(2*miu))*sum(max(-rd,0).^2);
end
end
penaltyQ1.m
if true
function [xmin,fxmin,it,errorx,rdxmin,rixmin]=penaltyQ1(funobj,x0)
kmax=100; %numero maximo de interacoes
errmax=10^(-7); % erro maximo entre duas solucoes
nx=length(x0);
[y,rd,ri]=funobj(x0)
nrd=length(rd);
nri=length(ri);
c=0.7;
miu=1;
k=0;
err=1;
while (k<kmax)&&(err>errmax) Q=@(x) funobj_qpm(funobj,miu,x); x1=fminsearch(Q,x0) % funcao objectivo passa a ser Q miu=c*miu; err=norm(x1-x0,2); %quadratic norm, (1 norm (1), infinite norm (inf)) x0=x1; k=k+1; [y,rd,ri]=funobj(x1); infxmin(k,1)=k; % linha k, faz 1 e sa? a linha k infxmin(k,2:(nx+1))=x1; %linha k, do 2 ao 3, sa? o xm infxmin(k,nx+2)=y; infxmin(k,(nx+3):(nx+nrd+2))=rd; infxmin(k,(nx+nrd+3):(nx+nrd+nri+2))=ri; infxmin(k,nx+nrd+nri+3)=err; end disp('it x* f(x*) rd(x*) ri(x*) erro') disp(infxmin); % solu??o encontrada [y,rd,ri]=funobj(x1); xmin=x1; %minimizante encontrado fxmin=y;% m?nimo encontrado it=k; errorx=err; rdxmin=rd; % valor na restri??o de desigualdade rixmin=ri; % valor na restri??o de igualdade % Como na fun??o temos 'funobj'.'RI' e 'RD' temos de definir fora. end
% Para executar %[xmin,fxmin,it,errorx,rdxmin,rixmin]=penaltyQ1(@func,[0;0;0;0;0]) %[xmin,fxmin,it,errorx,rdxmin,rixmin]=penaltyQ1(@func,[1;1;1;1;1]) %[xmin,fxmin,it,errorx,rdxmin,rixmin]=penaltyQ1(@func,[-1;-1;-1;-1;-1]) %[xmin,fxmin,it,errorx,rdxmin,rixmin]=penaltyQ1(@func,[0;1;0;1;0]) %[xmin,fxmin,it,errorx,rdxmin,rixmin]=penaltyQ1(@func,[2;0;3;0;0]) %[xmin,fxmin,it,errorx,rdxmin,rixmin]=penaltyQ1(@func,[rand(5,1)]) end
this last file penaltyQ1 is the file that have to execute and does not run
Jan on 9 Jan 2015
Please explain the occurring problems. "Does not run" is not useful. Do you get an error message?

Categories

Find more on Nonlinear Optimization in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!