Why is my newton method giving me a huge answer? What am i doing wrong?pls help
1 view (last 30 days)
Show older comments
x= sym('x');
p = 0.6;
F = x^5+4*x^2-1;
F_prime = diff(F);
N=0;
q = p - subs(F,x,p)/subs(F_prime,x,p);
while(N < 100 && not(abs(p-q) <= .00001 ))
N=N+1;
p = q;
q = p - ((subs(F,x,p))./(subs(F_prime,x,p)));
end
if(N == 100)
print("Too many iterations");
else
display(q)
end
0 Comments
Accepted Answer
John D'Errico
on 5 Sep 2019
When you have a problem like this, plot the function. In fact, plot the function anyway!
Next, it is a bad idea to use variables like p and q, then swaping them back and forth. Use variable names that make sense!!!!!
Next, don't keep things symbolic. That causes problems, because you don't know how to work with symbolic variables that well. Use functions instead.
I'll start with a symbolic function, then differentiate it in symbolic form. But then alllow matlabFunction to convert it to a function handle.
Finally, I cleaned up the logic in your code.
x = sym('x');
xcurrent = 0.5; % start point
xold = inf; % the first time through, you need to pass the while condition
F = x^3+4*x^2-10;
F_fun = matlabFunction(F);
fplot(F_fun)
hold on
plot(xcurrent,F_fun(xcurrent),'o')
F_prime = matlabFunction(diff(F));
N=0;
tolerance = 1.e-6;
while (N < 30) && (abs(xcurrent-xold) > tolerance )
N=N+1;
xold = xcurrent;
xcurrent = xold - F_fun(xold)./F_prime(xold);
plot(xcurrent,F_fun(xcurrent),'o')
end
if(N == 30)
print("Too many iterations");
else
display(xcurrent)
end
That should get the point across, that there is no need to use symbolic tools as heavily as you are.
More Answers (1)
Walter Roberson
on 5 Sep 2019
Your code is not giving you a huge answer: it is giving you a rational answer. You will want to use double() to convert it to floating point.
3 Comments
See Also
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!