Same code I can successfully run in R2021b but in R2016a it is showing Error using - Matrix dimensions must agree. What should I change to run this in R2016a?

function [facilities,R,Dij,E,C]=a(r1,r2,d1,d2,i,j)

tic

facilities=(1:i)';

R = randi([r1 r2],[i 2]);

R(:,2)=(R(:,1))+(R(:,2));

Dij=randi([d1 d2],[i,j]);

E=zeros(i,1);

E(:)=10^-6;

for m=1:i

for n=1:j

C = max((R(:,2)-Dij(:,:)),0)./max(max((R(:,2)-R(:,1)),(R(:,2)-Dij(:,:))),E(:,1));

n=n+1;

end

m=m+1;

end

toc

end

% Same code I can successfully run in R2021b but in R2016a it is showing Error using - Matrix dimensions must agree. What should I change to run this in R2016a?

[Jiri Hajek]

Hi, Matlab has given you much clearer indication about the error, but you have not posted it here, as well as data to reproduce your error.

• At which line of code the error occured?
• What was the value of variables entering the expression on that line?

[Sourasis Chattopadhyay]

• In line number 12 .
• [facilities,R,Dij,E,C]=a(10,20,30,40,5,3)

[Stephen23]

You can read more about implicit array expansion here:

https://blogs.mathworks.com/loren/2016/10/24/matlab-arithmetic-expands-in-r2016b/

The code throws an error on R2016a because it does not have implicit expansion.

Answer 1

[DGM]

R2016b introduced implicit array expansion. You're relying on that. The loops are also unnecessary.

i = 10;
j = 5;
r1 = 1;
r2 = 10;
d1 = 1;
d2 = 10;

facilities=(1:i)';
R = randi([r1 r2],[i 2]);
R(:,2)=(R(:,1))+(R(:,2));
Dij=randi([d1 d2],[i,j]);
E=zeros(i,1);
E(:)=10^-6;

num = max(bsxfun(@minus,R(:,2),Dij),0);
den = max((R(:,2)-R(:,1)),E(:,1)); % these terms agree
den = bsxfun(@max,bsxfun(@minus,R(:,2),Dij),den);
C = num./den;

I tested this in R2015b

[Sourasis Chattopadhyay]

Thank you so much for your valuable help.