Matrix dimensions do not agree

Hi everyone,

I am wondering if you can help me with this issue.

I have a differential equation code (dy/dt) that is solved using a solver. At the end I got the the dimension of y to be 2868x11.

There is also another parameter called xinf, with it being defined as

    xinf = 1/(1+exp(-(y(:,1)-Vx)/kx))

Vx and kx are both constant. At the end, xinf has the dimension of 1x2868.

The issue takes place when I need to calculate this formula:

Ix = gx*xinf.^3.*y(:,2).*(y(:,1)-Ex)

Ex is a constant.

I receive the error saying that the matrix dimensions do not agree. I tried transposing xinf, and although the code runs, it does not produce the output that is desired.

I was wondering is someone can help me solve this issue? Maybe another way to go about this? As I seem to have reached a dead end.

Many thanks in advance!

Answer 1

[James Tursa]

My guess is you meant that division in your xinf calculation to be element-wise. E.g., change the / to ./ as follows:

xinf = 1./(1+exp(-(y(:,1)-Vx)/kx))

[jnovv]

Thank you for your suggestion! Fixed the problem immediately :)

Answer 2

[the cyclist]

Just stating "it does not produce the output that is desired" is a bit vague. Can you be more specific?

Just a guess, though ... I bet you actually wanted

xinf = 1./(1+exp(-(y(:,1)-Vx)/kx))

rather than

xinf = 1/(1+exp(-(y(:,1)-Vx)/kx))

Notice that I used "./" in place of "/". That also solves your issue with needing to transpose, which gives me a dose of confidence that it was the right operation in the first place.

[jnovv]

Thank you very much for your help! The output needs to be in form of a graph that is similar to the data obtained from an experiment, which is quite tricky to explain. But your suggestion did fix it!