Getting error: Error using .* Matrix dimensions must agree. Please help. The code the given below.

 Nx = 100;Ny = 20;Nt = 4000;

 x = 0:10/Nx:10; y = -1:2/Ny:1; t = 0:100/Nt:100;

 Dx = 10/Nx;Dy = 2/Ny;Dt = 10/Nt;

 a = 0.001; Xd = a*Dt/(Dx^2); Yd = a*Dt/(Dy^2); c = Dt/(2*Dx); 

 T = zeros(Ny+1,Nx+1,Nt+1);

 T(1,1:end,:) = 0;

 for m = 1:Ny+1

    T(m,1,:) = (1 - (y(m).*y(m))).^2;

 end

 maxiter = 500;
 for k = 1:maxiter
     Tlast = T; % saving the last guess
     T(:,:,1) = Tlast(:,:,end); % initialize the scalar field at t = 0 to the last guess
     for i = 2:Nt+1
         T(2:end-1,2:end-1,i) = (1-(2*Xd)-(2*Yd))*T(2:end-1,2:end-1,i-1) + Xd*(T(2:end-1,3:end,i-1) + T(2:end-1,1:end-2,i-1)) - (c*1.5*(1 - (y(2:end).^2)).*(T(2:end-1,3:end,i-1) - T(2:end-1,1:end-2,i-1)) + Yd*T(3:end,2:end-1,i-1) + Yd*T(1:end-2,2:end-1,i-1);
         T(end,1:end,i) = (4*T(end-1,1:end,i) - T(end-2,1:end,i))/3;  % Top Wall (dT/dy = 0)
         T(1:end,end,i) = (4*T(1:end,end-1,i) - T(1:end,end-2,i))/3;  % Outlet (dT/dx = 0)
     end
     err(k) = max(abs(T(:)-Tlast(:)));  % finding the residual value between two successive iterations
     if err(k) < 1E-04
         break;     % stopping the solution is the residual value is small
     end
 end

The problem, I guess, lies in the formula for T(2:end-1,2:end-1,i).

[Justin]

In your giant line (seems to be missing a paren at the end in the above code?):

T(2:end-1,2:end-1,i) = (1-(2*Xd)-(2*Yd))*T(2:end-1,2:end-1,i-1) + Xd*(T(2:end-1,3:end,i-1) + T(2:end-1,1:end-2,i-1)) - (c*1.5*(1 - (y(2:end).^2)).*(T(2:end-1,3:end,i-1) - T(2:end-1,1:end-2,i-1)) + Yd*T(3:end,2:end-1,i-1) + Yd*T(1:end-2,2:end-1,i-1);

... each 'block' of math (for k=1) is producing a [19x99] double array, except this one:

 c*1.5*(1 - (y(2:end).^2))

which produces a [1x20] double, as y(2:end) is a [1x20] and c is a scalar.

Since I don't know the logic of what you're aiming for, I can't guess at what would properly fix that, but look around there and perhaps repost with the paren fixed on that line?

[Shashank]

Justin, thanks for your comment. But I checked that line again and I don't think I am missing a paren at the end.

The formula was changed to:

 c*1.5*(1 - (y(2:end-1).^2))

This is indeed, a [1x19] array. But it still gives me the same error. Can't find out why.

[Walter Roberson]

Justin is pointing out that you are trying to use .* to multiply a row vector by a 2D matrix. You might need to repmat() the row vector to make it the same size as the 2D matrix.

[Shashank]

Walter, thanks a lot. I'll try doing that.

[Shashank]

I tried creating a matrix like this but it did not work either.

 U(2:end-1,2:end-1,:) = 1 - ((y(2:end-1)).^2);

That's the error I am getting: Assignment has fewer non-singleton rhs dimensions than non-singleton subscripts

Any thoughts how it can be resolved?

[Shashank]

Thanks a lot peeps, for your valuable comments. I was able to solve it by defining a matrix as:

 U = zeros(Ny+1,Nx+1);
 for m = 1:Ny+1
    U(m,:,:) = 1 - ((y(m)).^2);
 end