Picture related allows to approximate the square root of any positive number.
I had an argument with two of my professors. They insisted that "sometimes" [math]R_i - R_{i+1}[/math] is a negative number
I argued that it only happens when you do [math]R_1 - R_{2}[/math] and never again
Who's right? Who's wrong?
>>8302521
What the fuck? Rn+1 will always be bigger than Rn are you retarded?
Like, what makes you think that making a number bigger will make its square root smaller?
>>8302530
huh? x=144
>>8302521
Let 0<x<1
easy enough to show, assuming
[math]R_i[/math] converges (which it does), just sub [math]R_ \infty[/math] into your relation,
and solve for [math]x[/math]
>>8302545
x=0.8
>>8302553
I don't know that much math
>>8302564
it's elementary Algebra:
[math]R_ \infty= \frac{1}{2} \left (R_ \infty+\frac{x}{R_ \infty} \right)[/math]
multiply by [math]2R{ \infty}[/math]:
[math]2R_{ \infty}^2=R_{ \infty}^2+x[/math]
subtract [math]R_{ \infty}^2[/math] from each side, and Bob's yer uncle
>>8302598
[math] R_ \infty= \frac{1}{2} \left (R_ \infty-1+\frac{x}{R_ \infty-1} \right) [/math] tho
>>8302598
And what does that prove?
>>8302621
wat
>>8302598
>infinity minus 1
keke, implying infinity is even a logically coherent mathematical entity
>>8302624
It proves that if the sequence converges, it must converge to sqrt(x), as this is the only fixed point.
It remains to be shown that the sequence converges.
>>8302521
If R_i > sqrt(x) then R_{i+1} < R_i.
because if R_i > sqrt(x) then 1 > x/(R_i)^2.
and 1 > 1/2(1 + x/(R_i)^2)
and so R_i > 1/2(R_i + x/R_i)
and so R_i > R_{i+1}.
Note for all x and y,
(y - sqrt(x))^2 > 0
and so y^2 - 2ysqrt(x) + x > 0
y^2 + x > 2ysqrt(x)
(y^2 + x)/2y > sqrt(x)
1/2(y + x/y) > sqrt(x)
R_i > sqrt(x), for all i != 1
so you are right OP. R_i < r_{i+1} only when i = 1.
>>8302708
>infinity
[math]R_{ \infty}[/math] is a number, you imbecile.
>>8302784
[math]If R_i > sqrt(x) then R_{i+1} < R_i. \\
because if R_i > sqrt(x) then 1 > x/(R_i)^2. \\
and 1 > 1/2(1 + x/(R_i)^2) \\
and so R_i > 1/2(R_i + x/R_i) \\
and so R_i > R_{i+1}. \\
Note for all x and y, \\
(y - sqrt(x))^2 > 0 \\
and so y^2 - 2ysqrt(x) + x > 0 \\
y^2 + x > 2ysqrt(x) \\
(y^2 + x)/2y > sqrt(x) \\
1/2(y + x/y) > sqrt(x) \\
R_i > sqrt(x), for all i != 1 \\
so you are right OP. R_i < r_{i+1} only when i = 1.[/math]