multiple outputs in symbolic calculation

I am trying to calculate the partial derivation of a function over different variables. Then I substitute the value of the variables in order to get a numeric value. The following code lien is used, in this particular case I want to derive a function F respect to a variable m, and x is an array containing the values of the variables:

 c1=double(subs(diff(F,m),{m,T1,ratio,I},x))

It works perfectly. However I want to iteratively repeat this computation. Therefore I have created a function that returns c1..c_n where n is the number of variables, the input of the function is the array x.

The idea is to assign the values of a matrix row to the array x and then introduce it in the function.

First I tried to make this step for one single iteration.

However i get a matrix n x m where m is the number of rows in the matrix that contains all the values that i want to evaluate. I dont know what can be due to or even if there is a direct relation between the number of columns of the matrix where i get my array from. Actually the values do not vary so much from one to another, but I would like to have a proper understanding of this behaviour.

For example, I introduce a vector

 x= [   1.2200   -0.5500         0    0.1500 ]

and I get

 c1 =

    1.1819
    1.1775
    1.1819
    1.1863
    1.1863
    1.1906
    1.1950
    1.1601
    1.1601
    1.1601
    1.1688
    1.1732
    1.1688
    1.1819
    1.1601
    1.1601
    1.1688
    1.1688
    1.1688
    1.1644
    1.1688
    1.1688
    1.1732
    1.1775
    1.1775
    1.1470
    1.1557
    1.1601
    1.1644
    1.1644
    1.1688
    1.1688
    1.1688
    1.1775
    1.1688
    1.1732
    1.1775
    1.1775
    1.1819
    1.1775
    1.1688
    1.1688
    1.1688
    1.1775
    1.1688
    1.1732
    1.1732
    1.1688
    1.1775
    1.1732
    1.1470
    1.1775
    1.1819
    1.1863
    1.1819
    1.1819
    1.1863

Has somebody experience such a behaviour?

Answer 1

[Walter Roberson]

Your F appears to resolve to a vector of length 57. When you subs() into a vector, you get one result for each element of the vector.

[Diego Fraguas Tejero]

Thanks for the fast answer Walter! It seems that the reason for this behavior is that I have a lot of vectors of size 57x1 double where I have the different variables stored. And therefore when I used the function double I got an output with that dimension.

However it always takes the row I want. If i take the next value, i got another set of 57 values of similar value

[Walter Roberson]

It is not double() that would lead to that: it is the subs() step, especially if what you are subs() into is a vector already.

[Diego Fraguas Tejero]

But if I try it separately in a new script, with a vector x that contains the values for the variables, there is no problem at all.

But when I have arrays in the workspace with a size 57x1 double then I get 57 outputs from the function. I do not think that is a coincidence.

Anyway, is there anything I can do to fix this problem? Perhaps an alternative way to solve my partial derivatives?

[Walter Roberson]

At the time of the problem, what does size(F) show?

[Diego Fraguas Tejero]

1 1. I just checked that diff() is already returning an output of 57 eq's

And by the way thanks for your attention :)

[Walter Roberson]

Could you confirm that your F is only length 1 but diff(F,m) is returning a vector of length 57? If so could you check size(m) at that point? If m were a symbolic vector of variable names then the result might not be as expected.

Also, is it possible that although your F is only length 1, that you defined it in terms of variables that were assigned to after they were used to construct F? What is size of subs(F) ?