what is wrong when using fzero

hi!

i would like to compute this:

function z=theta(m)

digits 250;

s=sym('s');
p=0;
q=0;

for i=0:m

      p=p+(factorial(2*m-i)*factorial(m)*s^i)/factorial(2*m)*factorial(m-
  i)*factorial(i);

      q=q+(factorial(2*m-i)*factorial(m)*(-s)^i)/factorial(2*m)*factorial(m-
  i)*factorial(i);

end;

r=p/q;

s=taylor(exp(-s)*r-1, s, 150); %taylor expansion

coeff=vpa(sym2poly(sym(s)),250); % c_i

x=sym('x');
sum1=0;
l=1;

for j=(2*m+1):(150+2*m)

      sum1=sum1+abs(coeff(l))*x^j;

      l=l+1;

end;

bound=vpa((-log(1-sum1))/x,250);

n=fzero(bound,3);

just the last part is important. i think there is something wrong with my use of fzero.. when i want to calculate this, i get this error:

>> theta(2)

??? Error using ==> fzero at 181

If FUN is a MATLAB object, it must have an feval method.

Error in ==> theta at 27

n=fzero(bound,3);

Answer 1

[Walter Roberson]

You cannot use fzero() on symbolic objects.

If you are going to use symbolic objects, use solve(), or use

   feval(symengine, 'numeric::solve', bound, 3)

[Vicky]

what if i want to use fzero; what should i change? thank you.

[Walter Roberson]

If you want to use fzero then you are going to have substantial difficulty in obtaining 250 digits of output.

[Vicky]

what about if i use feval(symengine, 'numeric::solve', bound, 3), like You wrote; im using Matlab version 7.9.0 (R2009b), so it doesnt work... how should i rewrite it?

[Vicky]

or if i can use fsolve, where there is an initial point, what i would like. How to form that...

[Walter Roberson]

Try

feval(symengine, 'numeric::solve', bound, x)

http://www.mathworks.com/help/symbolic/mupad_ref/numeric-solve.html

Answer 2

[Carlos]

You can use solve instead of fzero