ode15s XX must return a column vector

Hi,I am new to MATLAB. I need to calculate using ode15s, but I keep getting this error: XXX must return a column vector.

Here is my main script:

    dp0 = 2e-7;
    Np0 = 1e11;
    V0 = pi/6*dp0^3;
    C_bulk = 1.42e21*V0*Np0;
    c0 = [0;C_bulk];
    options = odeset('RelTol',1e-10);
    [t,c] = ode15s(@myfun,[0:120],c0,options,V0,C_bulk);
    plot(t,40.9*4.065e-14*338*c(1),'r.',
            t,40.9*4.065e-14*338*c(2),'b.');

The myfun.m file that includes the ode equations are below:

    function dcdt = myfun(t,c,V0,C_bulk)
    V = V0*c(2)/C_bulk;
    dp = (6*V/pi)^(1/3);
    dcdt(1) = 1.2e17*dp-c(1);
    dcdt(2) = c(1)-1.2e17*dp;

Please let me know why I get this error. Did I use the additional parameters (V0 and C_bulk) wrong? Thanks.

Answer 1

[James Tursa]

The "column vector" error can be resolved by changing the return variable into a column vector. E.g.,

    dcdt(1) = 1.2e17*dp-c(1);
    dcdt(2) = c(1)-1.2e17*dp;
    dcdt = dcdt(:); % <-- (:) notation results in column vector

When you build a vector piecemeal one index at a time like you do above, the default is for a row vector. Another way to solve this problem is to specify the column vector locations directly during the vector building process. E.g.,

    dcdt(1,1) = 1.2e17*dp-c(1);
    dcdt(2,1) = c(1)-1.2e17*dp;

And yet another way would be to pre-allocate the size of the return variable up front. E.g.,

    dcdt = zeros(2,1);
    dcdt(1) = 1.2e17*dp-c(1);
    dcdt(2) = c(1)-1.2e17*dp;

[H.O.Z.]

Thanks James, I have one more question regarding this function. If I have a local variable in myfun.m that changes with time. What is the best way to display this variable in the base workspace? Thank you.

[Walter Roberson]

You can assignin('base') the variable.