Infinite Square Roots!

Find the value of

1 17 16 1 17 16 1 \sqrt {1 - \sqrt {\dfrac {17}{16} - \sqrt {1 - \sqrt {\dfrac {17}{16} - \sqrt {1 - \ldots}}}}}

Assume, when taking the limit, that deep within the nested square roots we choose to begin with a “1” .


The answer is 0.5.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

4 solutions

Sharky Kesa
Jul 10, 2014

Let

x = 1 17 16 1 17 16 1 x = \sqrt {1 - \sqrt {\dfrac {17}{16} - \sqrt {1 - \sqrt {\dfrac {17}{16} - \sqrt {1 - \ldots}}}}}

Then

x = 1 17 16 x x = \sqrt {1 - \sqrt {\dfrac {17}{16} - x}} .

Simplifying, we get

x 2 = 1 17 16 x x^2 = 1 - \sqrt {\dfrac {17}{16} - x}

or

x 4 2 x 2 + 1 = 17 16 x x^4 - 2x^2 + 1 = \dfrac {17}{16} - x .

Simplifying further, we yield

x 4 2 x 2 + x 1 16 = 0 x^4 - 2x^2 + x - \dfrac {1}{16} = 0 .

Factorising this, we get

( 2 x 1 ) ( 8 x 3 + 4 x 2 14 x + 1 ) = 0 (2x - 1)(8x^3 + 4x^2 - 14x + 1) = 0 .

Therefore, either x = 1 2 x = \frac {1}{2} or x x is a root of 8 x 3 + 4 x 2 14 x + 1 = 0 8x^3 + 4x^2 - 14x + 1 = 0 . Solving this cubic equation, we get the roots x 1.62 , 0.073 x \approx -1.62, 0.073 and 1.05 1.05 . Since the answer to the problem has to be positive and less than one, the first and third roots are ruled out.

The only real roots left are 1 2 \frac {1}{2} and 0.073 0.073 . A double-check shows that neither was introduced in the squaring steps above. Both are correct solutions to

x = 1 17 16 x x = \sqrt {1 - \sqrt {\dfrac {17}{16} - x}} .

But x x clearly has one definite value, so which is it? A few iterations on a calculator (Sorry, the only way I thought I could do easily), starting with a "1" under the innermost radical, shows that x = 1 2 x = \frac {1}{2} is definitely the answer. Why?

The answer lies in the behaviour of f ( x ) = 1 17 16 x f(x) = \sqrt {1 - \sqrt {\frac {17}{16} - x}} near the points x = 1 2 x = \frac {1}{2} and x = α 0.073 x = \alpha \equiv 0.073 \ldots . The point x = 1 2 x = \frac {1}{2} is stable, in the sense that if we start with an x x value slightly different from 1 2 \frac {1}{2} , then f ( x ) f(x) will be closer to 1 2 \frac {1}{2} than x x was. The point x = α x = \alpha , on the other hand is unstable, in the sense that if we start with an x x value slightly different from α \alpha , then f ( x ) f(x) will be farther from α \alpha than x x was.

Said in another way, the slope of f ( x ) f(x) at x = 1 2 x = \frac {1}{2} is less than 1 in absolute value (it is in fact 2 3 \frac {2}{3} ), while the slope at x = α x = \alpha is greater than 1 in absolute value (it is approximately 3.4). What this means is that points tend to head toward 1 2 \frac {1}{2} , but away from α \alpha , under iteration by f ( x ) f(x) . In particular, if we start with the value 1, as we are supposed to do in this problem, then we will eventually get arbitrarily close to 1 2 \frac {1}{2} after many applications of f f . Therefore, 1 2 \frac {1}{2} is the correct answer.

There is more but I need to know how to upload pictures on solutions. This part makes it easier to understand the reasoning. I think this was my most hardcore solution I have made.

A really good, extensive analysis, Sharky. (If you manage to upload the pictures, even better.) I think that the critical point here is that "we start with the value 1". If the nesting were truly infinite that we would probably have to first deal with the issue of existence and uniqueness of a solution, (Real Analysis.... ugh). Also, I'm not sure if the fact that x = α x = \alpha represents an "unstable solution point" implies that it is not a solution, or if it just means that 1 2 \frac{1}{2} is the "better" solution. :)

If we replace 17 16 \frac{17}{16} with the variable y y and then evaluate the expression as y y approaches 1 1 , (from the right), then yet more interesting things happen. We appear to get potential limiting solutions of 0 , 1 0, 1 and ( ϕ 1 ) (\phi - 1) . I wonder if any of them are stable, though, and if the limit actually exists. Interesting stuff; there's more going on with your question than first meets the eye.

Brian Charlesworth - 6 years, 11 months ago
Daniel Liu
Jul 9, 2014

Let x x be equal to the expression.

We see that x = 1 17 16 x x=\sqrt{1-\sqrt{\dfrac{17}{16}-x}} .

Simplifying: x 2 = 1 17 16 x x^2=1-\sqrt{\dfrac{17}{16}-x}

Simplifying more: x 4 + 2 x 2 + 1 = 17 16 x x^4+2x^2+1=\dfrac{17}{16}-x

Even more: x 2 + 2 x 2 + x 1 16 = 0 x^2+2x^2+x-\dfrac{1}{16}=0

Using RRT: x = 1 2 x=\dfrac{1}{2} is a root.

Thus, we have that the answer is 0.5 \boxed{0.5} .

There is another root, about 0.073. You can test it out yourself. It seemingly fits the first equality, but why isn't it the answer? @Daniel Liu

Sharky Kesa - 6 years, 11 months ago

Log in to reply

Can you show how you are getting 0.073 0.073 as a root ? I am getting 0.5 , 0.92 0.5,-0.92 as the only real values.

Nishant Sharma - 6 years, 11 months ago

@Daniel Liu Sorry to be picky, but there should be a minus sign in front of the 2 x 2 2x^{2} term, in which case x = 1 2 x = \frac{1}{2} is indeed a root.

I agree with Sharky that there does also appear to be a solution of x = 0.0731827 x = 0.0731827 , (to 7 7 decimal places). There surely, however, must be a unique solution, so how can we rule out this second value? It satisfies Nishant's condition that x > 1 16 x \gt \frac{1}{16} , so what gives?

Brian Charlesworth - 6 years, 11 months ago

Log in to reply

Yes, please complete the solution.

Sharky Kesa - 6 years, 11 months ago
William Isoroku
Sep 12, 2015

x x is the expression.

So therefore

x 2 = 1 17 16 x x^2=1-\sqrt{\frac{17}{16}-x} \longrightarrow

16 x 4 1 + 16 x 32 x 2 = 0 16x^4-1+16x-32x^2=0

Factor and grouping:

( 4 x 2 + 1 ) ( 2 x + 1 ) ( 2 x 1 ) + 16 x ( 1 2 x ) = 0 (4x^2+1)(2x+1)(2x-1)+16x(1-2x)=0 \longrightarrow

( 2 x 1 ) ( ( 4 x 2 + 1 ) ( 2 x + 1 ) 16 x ) = 0 (2x-1)((4x^2+1)(2x+1)-16x)=0

One root is 1 2 \frac{1}{2} and the cubic I solved on calculator.

Now to choose the correct root is the last part. By close examination: the expression has to be greater than 0 0 so that means that 1 17 16 x > 0 1-\sqrt{\frac{17}{16}-x}>0 which limits the values to 1 16 < x < 17 16 \frac{1}{16}<x<\frac{17}{16}

Nishant Sharma
Jul 9, 2014

This solution is not complete

So many square roots. We'll find the domain first and then proceed with the equation.

Let the given expression equal x x . Then we have

x = 1 17 16 x ( i ) \displaystyle\,x=\sqrt{1-\sqrt{\frac{17}{16}-x}}---(i) .

Keeping in mind that square root function is true for positive values of arguments, we have

17 16 x > 0 \displaystyle\frac{17}{16}-x>0 and 1 17 16 x > 0 \displaystyle\,1-\sqrt{\frac{17}{16}-x}>0 . Both of them bound value of x x as

1 16 < x < 17 16 ( i i ) \displaystyle\frac{1}{16}<x<\frac{17}{16}---(ii)

Now taking ( i ) (i) we have

1 x 2 = 17 16 x \displaystyle\,1-x^2=\sqrt{\frac{17}{16}-x}

Squaring both sides

x 4 2 x 2 + x 1 16 = 0 \displaystyle\,x^4-2x^2+x-\frac{1}{16}=0

Now using Rational Root Theorem we factor the above biquadratic as

( x 1 2 ) ( x 3 + x 2 2 7 x 4 + 1 8 ) = 0 \displaystyle\left(x-\frac{1}{2}\right)\left(x^3+\frac{x^2}{2}-\frac{7x}{4}+\frac{1}{8}\right)=0 .

We see that the second polynomial doesn't have a positive root( I couldn’t help myself from using W|A for this \textit{I couldn't help myself from using W|A for this} ), so we conclude that x = 1 2 = 0.5 x=\frac{1}{2}=\boxed{0.5} is a root which indeed satisfies ( i ) (i) and ( i i ) (ii) .

So does a number roughly 0.073. Solution isn't fully complete.

Sharky Kesa - 6 years, 11 months ago

Log in to reply

Can't follow where does this value come from ? Could you post your solution please ?

Nishant Sharma - 6 years, 11 months ago

Log in to reply

One way to find the value is to plot the first equation. Two intersections are found, 0.5 and 0.073. I'll post my solution.

Sharky Kesa - 6 years, 11 months ago

Log in to reply

@Sharky Kesa Is there a closed form for the same ?

Nishant Sharma - 6 years, 11 months ago

Solution as requested.

Sharky Kesa - 6 years, 11 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...