Does This Sequence Converge to the Fixed Point?

GregA
Messages
210
Reaction score
0
[SOLVED] Proof of convergence

Homework Statement


Let \alpha[/tex] be a fixed point of x = g(x) and let (x_n) be the sequence generated by the fixed point iteration scheme. Using the first two terms of the Taylor series for<br /> g(x) about \alpha[/tex] we can get an approximation for g(x_n):&lt;br /&gt; g(x_n) = g(\alpha) + (x_n - \alpha)g&amp;amp;#039;(\alpha)[/tex]&amp;lt;br /&amp;gt; (Assuming the terms in the sequence are close to α we have neglected non-&amp;lt;br /&amp;gt; linear terms.)&amp;lt;br /&amp;gt; &amp;lt;br /&amp;gt; Show that |x_n - \alpha| = |x_0 - \alpha||g&amp;amp;amp;#039;(\alpha)|^{n}[/tex] for all n \geq 1 hence show that the sequence (x_n) converges to \alpha[/tex] for g&amp;amp;amp;amp;amp;#039;(x)&amp;amp;amp;amp;amp;lt; 1&amp;amp;amp;lt;br /&amp;amp;amp;gt; &amp;amp;amp;lt;h2&amp;amp;amp;gt;Homework Equations&amp;amp;amp;lt;/h2&amp;amp;amp;gt;&amp;amp;amp;lt;br /&amp;amp;amp;gt; &amp;amp;amp;lt;h2&amp;amp;amp;gt;The Attempt at a Solution&amp;amp;amp;lt;/h2&amp;amp;amp;gt;&amp;amp;amp;lt;br /&amp;amp;amp;gt; I&amp;amp;amp;amp;#039;m not so worried about the proving second part (looks obvious if the first part is true) but before trying to prove the first I want to try a couple of examples and see what&amp;amp;amp;amp;#039;s happening if I can.&amp;amp;amp;lt;br /&amp;amp;amp;gt; If I suppose that my function g(x) = \sqrt{x+1} then the value of a fixed point \alpha = \frac{1+\sqrt{5}}{2}[/tex]&amp;amp;amp;amp;lt;br /&amp;amp;amp;amp;gt; Now if I let x_0 = 2 and set n = 1 then x_1 = \sqrt{2+1} and I am under the impression that I can now show:&amp;amp;amp;amp;lt;br /&amp;amp;amp;amp;gt; |\sqrt{3} - \alpha| = |2 -\alpha||\frac{1}{2}(\frac{1}{\sqrt{\alpha+1}})|[/tex]&amp;amp;amp;amp;amp;lt;br /&amp;amp;amp;amp;amp;gt; \Rightarrow |\sqrt{3} - \frac{1+\sqrt{5}}{2}| = |2 - \frac{1+\sqrt{5}}{2}||\frac{1}{2}(\frac{1}{\sqrt{\frac{1+\sqrt{5}}{2}+1}})|&amp;amp;amp;amp;amp;lt;br /&amp;amp;amp;amp;amp;gt; But this is false!...How am I misinterpreting the given statement or what have I done wrong?&amp;amp;amp;amp;amp;lt;br /&amp;amp;amp;amp;amp;gt; Please don&amp;amp;amp;amp;amp;amp;#039;t prove the problem for me.
 
Last edited:
Physics news on Phys.org
You are right, it is not true- it is only approximately true. After all, you neglected powers higher than the first in the Taylor expansion.

Proving |x_n- \alpha|=|x_0- \alpha||g&#039;(\alpha)^n, to first order, by induction, is pretty easy if you replace g(x_n) and g(\alpha) in g(x_n) = g(\alpha) + (x_n - \alpha)g&#039;(\alpha) by x_{n+1} and \alpha.
 
Last edited by a moderator:
Cheers for that HallsofIvy, it was the presence of an equals sign, not approximately equals that was throwing me; I couldn't see why it should have been true :redface:
 
Prove $$\int\limits_0^{\sqrt2/4}\frac{1}{\sqrt{x-x^2}}\arcsin\sqrt{\frac{(x-1)\left(x-1+x\sqrt{9-16x}\right)}{1-2x}} \, \mathrm dx = \frac{\pi^2}{8}.$$ Let $$I = \int\limits_0^{\sqrt 2 / 4}\frac{1}{\sqrt{x-x^2}}\arcsin\sqrt{\frac{(x-1)\left(x-1+x\sqrt{9-16x}\right)}{1-2x}} \, \mathrm dx. \tag{1}$$ The representation integral of ##\arcsin## is $$\arcsin u = \int\limits_{0}^{1} \frac{\mathrm dt}{\sqrt{1-t^2}}, \qquad 0 \leqslant u \leqslant 1.$$ Plugging identity above into ##(1)## with ##u...
Back
Top