How Does Fixed Point Iteration Converge with Nested Square Roots?

Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
3 replies · 2K views
math8
Messages
143
Reaction score
0
Let p>0 and [tex]x = \sqrt{p+\sqrt{p+\sqrt{p+ \cdots }}}[/tex] , where all the square roots are positive. Design a fixed point iteration [tex]x_{n+1} = F (x_{n})[/tex] with some F which has x as a fixed point. We prove that the fixed point iteration converges for all choices of initial guesses greater than -p+1/4.



Let [tex]x_{n+1}=F(x_{n})= \sqrt{p+x_{n}}[/tex] so x is a fixed point for F since F(x)=x.
Now let [tex]g(x)=\sqrt{p+x}[/tex].
We have [tex]g'(x)=\frac{1}{2 \sqrt{p+x} }[/tex]


I can see that for [tex]x > -p + 1/4[/tex], we have that g'(x) <1.

From there I am not sure how to proceed to obtain convergence for [tex]x_{0} > -p +\frac{1}{4}[/tex] .
 
Last edited:
Physics news on Phys.org
It (g(x)) must be continuously differentiable on a closed interval [a,b], and g([a,b]) C[a,b].
Also, max {|g' (x)|: x in [a,b] } < 1.

Then the iterations converge to the unique fixed point for any initial guess x_0 in [a,b].

I can see that for x > -p + 1/4 , we have that g'(x) < 1. So can I say that max {|g' (x)|: x> -p + 1/4 } < 1 ? Also, I am not quite sure how to deal with the fact that the interval that I have in this case is an open interval (-p + 1/4 , infty).
 
You can look at the set [itex][a,\infty)[/tex] in two ways:<br /> <br /> 1. As the limit of a closed interval as the upper bound goes to infinity:<br /> <br /> [tex][a,\infty) = \lim_{b\to\infty} [a,b][/tex]2. As the union of all closed intervals [a,b]:<br /> <br /> [tex][a,\infty) = \bigcup_{b>a}\,[a,b][/tex]Either way, you can use the results you have already obtained to show that the fixed point iteration converges.[/itex]