I would like to minimize FT^-1(Ax-B) using lsqr. A is a huge sparse matrix and I defined it using 3 vectors:
RA(L) is the Lth nonzero of A, counting across row 1, then across row 2, and so on;
JA(L) is the column in which the Lth nonzero of A lies;
NA(I) is the number of nonzero coefficients in the Ith row of A.
I computed Ax using a user-defined nested function APROD that I'm trying to pass as a function handle in lsqr. However, it returns an error message: Too many input arguments.
This is my code:
function x = PROVA
%RA, NA, JA, m are defined here;
mode=1;
x=lsqr(@APROD,B,atol,itnlim);
end
function result=APROD(x1)
%if mode=1 computes y=y+Ax
%if mode=2 computes x=x+A_(transpose)y
%A is stored in RA,JA,NA by rows
L2=0;
if mode==1
result=zeros(m,1);
for c=1:m
sum=0;
L1=L2+1;
L2=L2+NA(c);
for L=L1:L2
J=JA(L);
sum=sum+RA(L)*x1(J);
end
result(c)=result(c)+sum;
end
end
if mode==2
result=zeros(n,1);
for c=1:m
Yc=y(c);
L1=L2+1;
L2=L2+NA(c);
for L=L1:L2
J=JA(L);
result(J)=result(J)+RA(L)*Yc;
end
end
end
end
This is the error message:
Error using iterapp (line 59)
user supplied function ==> APROD failed with the following error:
Too many input arguments.
Error in lsqr (line 190)
v = iterapp('mtimes',afun,atype,afcnstr,u,varargin{:},'transp');
Error in PROVA (line 97)
x=lsqr(@APROD,B,atol,itnlim);
I don't know what I did wrong, it is my first time with function handle, I read other questions about it but I didn't manage to solve my problem. Thank you for your help!
The function handle should take 2 parameters: first one is the solution approx vector, and the second is a char that can be 'notransp'
or 'transp'
:
function result=APROD(x1, mode)
L2=0;
switch mode
case 'notransp'
result = zeros(m,1);
for c=1:m
sum=0;
L1=L2+1;
L2=L2+NA(c);
for L=L1:L2
J=JA(L);
sum=sum+RA(L)*x1(J);
end;
result(c)=result(c)+sum;
end;
case 'transp'
result = zeros(n,1);
for c=1:m
Yc=y(c);
L1=L2+1;
L2=L2+NA(c);
for L=L1:L2
J=JA(L);
result(J)=result(J)+RA(L)*Yc;
end
end
end
end
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.