How do I apply Newton's method to this problem?

I need to write a MATLAB program that uses the Newton's Method to find the root of this function:

x^2-e^2-2=0
Xo=-2

Root has to be accurate to the 10^-6

I know how to do the Newton's method. The MATLAB file requires three files: function, function derivative, and newton function. It also requires using the FEVAL method to find the solution to the function derivative and function.

[Sean de Wolski]

It won't require feval if you do it properly...

Anyway, post what you've done so far because no one here will do your homework for you.

[Ashley Dunn]

1st function file

function fun(x)

x^2-exp^x+2; %#ok<VUNUS>

end

2nd function file

function functionderivative(x)

2*x-exp(x);

end

3rd function file

function result = newtroot (fun, functionderivative, x)

delta = tol = sqrt (eps);

maxit = 50;

fx = feval (fun, x);

deriv=feval(functionderivative, x);

for i = 1:maxit

if (abs (fx) < tol)

result = x;

return;

else

fx_new = feval (fun, x + delta);

x = x - fx / deriv;

fx = fx_new;

endif

endfor

***I want to use feval...that was the objective of the problem...the error says that there are too many output arguments for the "fun" function.

[Paulo Silva]

function [y]=fun(x)

y=x^2-exp^x+2;

end

[Sean de Wolski]

or

fun = @(x)x^2-exp^x+2;

Anyway, no reason for feval, tell your teacher/professor to use function handles!

[Ashley Dunn]

Lol....I know how to input a function into an m-file....how am I supposed to use feval to Newtonian method up the original function??

Answer 1

[Walter Roberson]

"Lol....I know how to input a function into an m-file" ... The "fun" function you constructed has no return value defined, which is what the error message was complaining about. Paulo shows how to define it with an output argument. Do likewise for your functionderivative function.

Now, when you invoke your newtroot, be sure to pass @fun or 'fun' as the first argument rather than fun itself without any adornment. Likewise, in newtroot, do not feval(functionderivative,x) and instead feval(@functionderivative,x) or feval('functionderivative',x) . What you have now will evaluate functionderivative with no arguments, and expects that function to return a value that is the name string or function handle to be evaluated.