Proving the Inequality in Newton's Square Root Method

AI Thread Summary
The discussion focuses on proving the inequality |sqrt(a) - e| < |a/e - e| in the context of Newton's Square Root Method. Participants analyze the relationship between the approximated value e and the actual square root of a, considering different scenarios where e is either less than, equal to, or greater than sqrt(a). The conversation highlights the importance of understanding the implications of these inequalities and how they guide the iterative process of Newton's method. Participants express confusion about the steps involved in the proof and seek clarification on the logic behind the inequalities. Ultimately, the goal is to establish a clear mathematical relationship that validates the convergence of e to sqrt(a).
Bonaparte
Messages
26
Reaction score
0

Homework Statement



Let e be the number close to sqrt(a) by Newtons Method (That is picking a number, diving a by it, and taking their average, divide a by average, get a number, find their average, so on). Using |e<sqrt(a)+e|
prove that if |a/e-e|<1/10
then |sqrt(a)-e|<1/10

Note that e is using the Newtons method a few times, not necessarily infinity, for any number of times. Also this is about positive integers, and 0 only, root and a.

Homework Equations

The Attempt at a Solution


So were trying to prove the second one smaller then first (I think), that is:
|sqrt(a)-e|<|a/e-e|
sqrt(a)-e<a/e-e (since both are positive, as using the given inequality subtract e from both sides) so sqrt(a)<a/e
e*sqrt(a)<a,
but e is not necessarily smaller then sqrt a, what am I missing?
 
Last edited:
Physics news on Phys.org
e =>0 and e<sqr(a) with a>0 right?

if e == sqr(a) then you'd have e*e = a and hence a < a which is wrong
 
Sorry, i had a typo, its a +, not a -,| e< sqrt(a)+e|
 
Bonaparte said:
Sorry, i had a typo, its a +, not a -,| e< sqrt(a)+e|

Has that resolved your problem or are you still stuck?
 
Very stuck PLEASE HELP
 
Bonaparte said:
Very stuck PLEASE HELP

Be more specific where exactly and why? sometimes in just explaining and thinking about it the answer will come.
 
So were trying to prove the second one smaller then first (I think), that is:
|sqrt(a)-e|<|a/e-e|
sqrt(a)-e<a/e-e (since both are positive, as using the given inequality subtract e from both sides) so sqrt(a)<a/e
e*sqrt(a)<a,
but e is not necessarily smaller then sqrt a, what am I missing?
There :)
 
To find \sqrt{a}, we choose some starting value, e, and calculate a/e. There are three possibilities:

1) e= \sqrt{a}. Then e^2= a so that e= a/e. We get the same number again and so know that we are done.

2) e&lt; \sqrt{a}. Then multiplying both sides by \sqrt{a}, e\sqrt{a}&lt; a and \sqrt{a}&lt; a/e. That is, e&lt; \sqrt{a}&lt; a/e. The square root of a is somewhere e and a/e and the midpoint, (e+ a/e)/2, is as good a place to look as any.

3) e&gt; \sqrt{a}. Then multiplying both sides by \sqrt{a}, e\sqrt{a}&gt; a and \sqrt{a}&gt; a/e. That is, e&gt; \sqrt{a}&gt; a/e. The square root of a is somewhere e and a/e and the midpoint, (e+ a/e)/2, is as good a place to look as any.
 

Similar threads

Back
Top