1
$\begingroup$

This question already has an answer here:

I am trying to show that $s_{n+1} = \frac{1}{2} \left(s_n + \frac{3}{s_n}\right)$ is a monotonic decreasing sequence for $n \ge 2$. Currently, my approach using induction is stuck because of the $\mathbf{s_n}$ term appearing in the denominator.

What I have so far:

Calculating a few terms: $s_1 = 1$, $s_2 = 2$, $s_3 = \frac{7}{4} < s_2$. For this to be monotonic decreasing, we have to show $s_{k+2} < s_{k+1}$.

For the base case, we assume for $k$ we have $s_{k+1} < s_k.$

The inductive step: to show $s_{k+2} < s_{k+1}$, i.e. to show $$\color{grey}{s_{k+2} = }\frac{1}{2} \left(s_{k+1} + \frac{3}{s_{k+1}}\right) < \frac{1}{2} \left(s_{k} + \frac{3}{s_{k}}\right) \color{grey}{= {s_{k+1}}_.}$$

In order to show $\dfrac{1}{2} \left(s_{k+1} + \frac{3}{s_{k+1}}\right) < \dfrac{1}{2} \left(s_{k} + \frac{3}{s_{k}}\right)$, I am stuck:

Given, $s_{k+1} < s_k$, I cannot then say $s_{k+1} + \mathbf{\frac{3}{s_{k+1}}} < s_k + \mathbf{\frac{3}{s_{k}}}$, because the $s_k, s_{k+1}$ terms which appears in the denominator may reverse the inequality. Any pointers on how to proceed further?

Disclaimer: I am revising real analysis on my own from the Kenneth Ross book, this is not strictly homework.

$\endgroup$

marked as duplicate by rtybase, Namaste, JonMark Perry, Claude Leibovici, Parcly Taxel Feb 28 '18 at 1:34

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • 1
    $\begingroup$ If $s_1=1$ then $s_2=2$. How's this sequence a monotonically decreasing sequence then? $\endgroup$ – Math Lover Feb 26 '18 at 20:34
  • $\begingroup$ You are applying the Babylonian method for computing $\sqrt{3}$, which is equivalent to Newton's method applied to $f(x)=x^2-3$. $\endgroup$ – Jack D'Aurizio Feb 26 '18 at 20:42
  • 1
    $\begingroup$ This is Newton-Raphson method to solve the equation $x^2-3=0$. $\endgroup$ – hamam_Abdallah Feb 26 '18 at 20:44
  • $\begingroup$ @MathLover corrected the question in response to your comment $\endgroup$ – AruniRC Feb 26 '18 at 20:44
1
$\begingroup$

Presuming $s_1 > 0$, by AM-GM inequality, $$s_{n} = \frac{s_{n-1}+\frac{3}{s_{n-1}}}{2} \ge \sqrt{3}$$ for $n \ge 2$. Also, $$s_{n+1}-s_{n} = \frac{1}{2}\left(\frac{3}{s_n}-s_n\right)\le 0 \iff s_{n} \ge \sqrt{3}$$ for $n \ge 2$.

$\endgroup$
1
$\begingroup$

Assume $s>0$ and notice that for $s\ne\sqrt3$

$$\frac12\left(s+\frac3s\right)>\sqrt3$$ as can be established by computing the minium.

So for $n>1$, we have

$$s_n>\sqrt3\implies s_n^2>3\implies s_n>\frac3{s_n}\implies s_n>\frac12\left(s_n+\frac3{s_n}\right)=s_{n+1}.$$


At the same time this establishes that the sequence is decreasing and bounded below by $\sqrt3$, thus is convergent.

$\endgroup$
0
$\begingroup$

The hint.

Use for all $n\geq2$ $$s_{n+1}-\sqrt3=\frac{(s_n-\sqrt3)^2}{2s_n}>0$$ and $$s_{n+1}-s_n=\frac{3-s_n^2}{2s_n}<0.$$

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.