Output: Long numbers for no reason

Hi,

it's me with another question - sorry! ;-) I have encountered this problem for a while now and it's more and more annoying so I decided to ask here: Often, Matlab doesn't calculate stuff "until the end". For example: I get the info that the solution for a random task is:

ans=-(1152921504606846976*sin(7821430514075775/72057594037927936) - 368769787030653356*cos(7821430514075775/72057594037927936) + 46928583084454650*cos(54750013598530425/72057594037927936) + (23464291542227325*sin(54750013598530425/72057594037927936))/2 + 288230376151711744*cos(23464291542227325/36028797018963968) - 1152921504606846976*sin(23464291542227325/36028797018963968) - 241301793067257094*cos(39107152570378875/72057594037927936) + (2329307300755921277*sin(39107152570378875/72057594037927936))/2 + 274912620861744056)/(50*(576460752303423488*cos(7821430514075775/72057594037927936) + 23464291542227325*sin(7821430514075775/72057594037927936) - 576460752303423488))

This, of course, doesn't help me much. I have to copy and paste the whole expression into the command window again and end up getting the final value of approx. 0.012521. I guessed that it has something to do with the numeric format but various tries haven't changed anything, it's the same in every format. Of course I used Google but that didn't lead to anything in this case. So I hope someone here might have an answer for me?

Answer 1

[Steven Lord]

When you perform computations using sym objects, MATLAB will not approximate the result of those computations as a double precision result normally. In order to approximate the results you need to either do something that requires the results to be double precision (like assigning the variable into an element of a double array) or explicitly tell MATLAB to approximate it using the double or vpa functions.

two = sym(2);
sqrt2 = sqrt(two) % displayed symbolically
x = 0;
x(1) = sqrt2
s2 = double(sqrt2)
s2_2 = vpa(sqrt2)

Note that sqrt2 is displayed as 2^(1/2) but x, s2, and s2_2 all display that quantity to a certain number of decimal places. The format function will not affect the display of sqrt2 or s2_2 but will affect the display of x or s2.

[Backtobasics]

Thank you very much! :) A bit annoying that there's no preference to tell Matlab to do it automatically but at least I can avoid the copy+paste. ;-)

Is there any way to apply double() to vectors as well? I get the error "DOUBLE cannot convert the input expression into a double array."

[Walter Roberson]

If you have an symbolic expression (including a scalar) that includes a symbolic variable, then it is not possible to convert that expression to double. For example if you had

syms x
f = sin(x);

then double(f) would fail, as f contains the unresolved symbolic variable x.

There is no problem applying double() to symbolic arrays of any dimension provided that the expressions do not contain symbolic variables.

(Except, that is, for the cases where the numeric equivalent cannot be calculated, such as a request to integrate an expression that is too jumpy to do numeric integration on.)