fsolve problem

Hi I am having two problems: I--> I used fsolve in two ways. one by declaring the function inside fsolve and second by making a function. Both give different answers. Second method gives complex solution which is not possible. II--> For some equations fsolve is complex answers. Can i by some method limit fsolve to give real solution.

Answer 1

[Jesús Zambrano]

You can try with the function solve.

For example, to force getting real solutions:

syms x
eqn = x^5 == 3125;
solve(eqn, x) % gives all the solutions
solve(eqn, x, 'Real', true) % for ce to give the real solutions

[madhan ravi]

but solve() is used in the answer

[Jesús Zambrano]

I meant: function solve.

Answer 2

[Walter Roberson]

It is not possible to set bounds for fsolve() . You can set the FunValCheck option for fsolve() to tell it to check each iteration for nan and inf and complex values, but that will cause fsolve() to error, not to somehow avoid or work around the problem.

If you just happened to be working with a single function with a single variable, then fzero() permits a range to be passed for x0.

If you have the symbolic toolbox, then vpasolve() permits you to pass bounds for multiple variables.