Decimal numbers in a triple for loop

Hi all, I want to run the equation for a range of q variables at an increment such as

q = 0.1:0.1:1;

I need to use a for loop, where actually I include 2 additional for loops:

a & b are 2 cells.

M=10;

for i=1:length(q)

          for i1=1:M
          for i2=1:M
              n_pm(i)=n_pm(i)+exp(-1i*q(i)*(i1-i2))*a{i1}*b{i2};
          end
          end
          n_pmT(i)=n_pm(i)/(2*M)^2;

end

This code however gives me either the error 'Subscript indices must either be real positive integers or logicals' or 'Index exceeds matrix dimensions', depending on how I define q in the loop.

Looking forward for some help!

Thank you in advance.

[James Tursa]

It would help if you copy & pasted the entire error message for us, including the offending line.

Answer 1

[James Tursa]

Have you inadvertently created a variable with the name 'exp' in your workspace? If so, clear it and don't use that name for a variable.

[Auryn_]

Hi, no, that is actually e^(x).

[James Tursa]

Type this at the command prompt:

dbstop if error

Then run your code. When the error occurs the code will pause with all variables intact. Examine them to see what is going on.

The exp( ) still seems like the likely culprit ...

[Auryn_]

Hi,

the error 'Index exceeds matrix dimensions' appears even if I use:

q = 0.1:0.1:1;

for i=1:length(q)

        for i1=1:M

        for i2=1:M

            n_pm(i)=n_pm(i)+q(i)*(i1-i2)*a{i1}*b{i2};

        end

        end

        n_pmT(i)=n_pm(i)/(2*M)^2;

end

[James Tursa]

This error:

'Subscript indices must either be real positive integers or logicals' or 'Index exceeds matrix dimensions'

was probably caused by a variable named 'exp' in your workspace.

This error:

'Index exceeds matrix dimensions'

is different, and means exactly what it says. Did you use the dbstop debugging tip I gave you? When the code paused, what were these values:

i
i1
i2
numel(q)
numel(n_pm)
numel(n_pmT)
numel(a)
numel(b)

Answer 2

[Auryn_]

Hi,

now it seems to work!

 M=10;

 q = 0.1:0.1:1;

 n_pmT=zeros(length(q),1);

 n_pm=zeros(M,M,length(q));

 for i=1:length(q)

        for i1=1:M

            for i2=1:M

                n_pm=n_pm+exp(-1i*q(i)*(i1-i2))*a{i1}*b{i2};

            end

        end

    n_pmT(i)=n_pm(i)/(2*M)^2;

 end

[James Tursa]

Well, it may run but I have my doubts that this code actually does what you want it to. Inside your inner loop you have n_pm=n_pm+etc. Do you realize that the etc part is added to all of the elements of n_pm at each iteration? Is that what you really want?

[Auryn_]

Hi James, I want a sum over the indices i1 and i2. Is that correct in the code? Thanks again!