cos cos cos ...

Calculus Level 4

Let f ( x ) = cos x f(x)=\cos x . For all real numbers x x , find lim n ( f f f n times ) ( x ) . \lim_{n \rightarrow \infty}(\underbrace{f \circ f \circ \cdots \circ f}_{n\text{ times}})(x).

Express your answer to 5 decimal places.


The answer is 0.739085.

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.

5 solutions

Abhishek Sinha
Sep 29, 2014

Note that, for any x R x \in \mathbb{R} , we have cos ( x ) [ 1 , 1 ] \cos(x) \in [-1,1] and hence cos ( cos ( x ) ) = [ cos ( 1 ) , 1 ] [ 0.54 , 1 ] \cos(\cos(x))=[\cos(1),1]\approx [0.54,1] . Denote the compact set [ cos ( 1 ) , 1 ] [\cos(1),1] by S. Since the function f ( ) = cos ( ) f(\cdot)=\cos(\cdot) is a non-increasing function in the interval S S , it is evident that f f maps all points in S S onto S S . Now consider the mapping f : S S f :S\to S . For any x S , f ( x ) = sin ( x ) sin ( 1 ) < 1 x \in S, |f'(x)|=\sin(x)\leq \sin(1) < 1 . Hence the mapping f : S S f:S\to S is a contraction . Now from a direct application of Banach's fixed point theorem , we conclude that the limit in the question exists and is equal to the unique fixed point of the equation cos ( x ) = x \cos(x)=x in the set S S , which is approximately 0.739 0.739 .

Can you explain it in simple manner, did'nt understood your solution

U Z - 6 years, 5 months ago

Too good, too good. Nicely done, sir.

L N - 6 years, 8 months ago

oh it's radians i thought it was in degrees :(

Rindell Mabunga - 6 years, 8 months ago

Log in to reply

The problem I had created originally set x x in degrees, and later on it was changed to radians. See comments below. :)

Jaydee Lucero - 6 years, 8 months ago

Well, what I did was press the cosine button on my calculator about 1000 times and press a random number, then repeat, and both times it gave me the exact same number, . 99985 .99985 .

Do I need to press that button 1,000,000 times to see the difference or what? Lol

John M. - 6 years, 5 months ago

This argument can be made more rigorous but here is the general point.

Let x = ( f f f ) ( t ) x=(f\circ f \circ f \circ \cdots)(t) be the value denoting an infinite convolution of the function f f evaluated at t t . Applying f f to both sides preserves the RHS and gives us f ( x ) = x f(x)=x as a functional equation.

Now for the case f = cos f=\cos we are solving for cos ( x ) = x \cos(x)=x which is not doable explicitly so just grab a calculator and take cos of 1 50 times until the fifth decimal place stops changing or use a CAS. I don't know of anyway to solve such an equation so I will try and have a look around about it.

The answer is 0.739 \boxed{0.739} .

I know a method can solve this equation, it may be long but useful, using Newton's Method .

Let f ( x ) = x cos x f(x)=x-\cos x , and let x 0 = 0.5 x_0 = 0.5 be our first assumption for the root of f f .

Applying the formula several times ( Here, I think it wouldn't take much time, and the fifth decimal will be stable ) :

x n = x n 1 f ( x ) f ( x ) \displaystyle x_n=x_{n-1} - \frac{f(x)}{f'(x)} .

Hasan Kassim - 6 years, 5 months ago

There is no trivial way to solve such an equation: http://math.stackexchange.com/questions/46934/what-is-the-solution-of-cosx-x

L N - 6 years, 8 months ago
Akeel Howell
Apr 22, 2017

Similar solution posted here .

I tried to answer this sum logically instead of working it out.Now the range of cosine function is between -1 and 1.So if we arbitrarily choose a suitable angle like ( say 0 degrees or 60 degrees) we would see that the argument goes on diminishing and finally tends to 0. As we extrapulate n to infinity the argument would become 0 and as a result cosine of 0 degrees is equal to one and hence the result

Your argument is fallacious.

The answer of 1 is marked correct because it is close to the numerical answer of 0.9998.

I have updated the problem phrasing.

Calvin Lin Staff - 6 years, 8 months ago
L N
Sep 28, 2014

So: Let: S = lim n ( f f f ) S = \lim_{n \to \infty} (f \circ f \circ \ldots \circ f) Then:

S ( x ) = y S(x) = y and cos ( S ( x ) ) = y \cos(S(x)) = y So, substitute the first equation into the second:

c o s ( y ) = y cos(y) = y

Using like Taylor Series ( x = cos ( x ) x = \cos(x) has no closed form), or better yet just Wolfram Alpha for a value, you get:

0.99985 0.99985

(remember to tell Wolfram Alpha degrees, or else you'll get 0.7 something, like I first did)

I hope I did this correctly.

I've updated the problem phrasing to remove that the function is in degrees, as that is not standard in calculus. Also, there were many people who entered 1, for the wrong reason, and were marked correct because the value was too close.

@Kenny Lau Note that your dispute was not valid (even though you are now marked correct) because the question initially mentioned that the calculations were done in degrees. I have since edited the phrasing.

Calvin Lin Staff - 6 years, 8 months ago

Log in to reply

Oops, I'm sorry.

Kenny Lau - 6 years, 8 months ago

Maybe Brilliant should allow people to determine the sensitivity of decimal answers? (Or do they already? I haven't posted any questions on here...) Also, small mistake in my answer: Taylor Series I don't think will get you anywhere in particular if you try doing it by hand, a more appropriate choice would be to do something like Newton's method. At least, I'm pretty sure, haven't tested Newton's method either, haha.

L N - 6 years, 8 months ago

Log in to reply

Right now, the sensitivity of decimal answers is standardized at ± 2 % \pm 2\% . I think that setting own sensitivity would likely result in a lot more confusion / debate over the actual range instead, which is not constructive.

There isn't an analytic solution to this problem. We can get close approximations of the value, but likely not the exact value itself.

Calvin Lin Staff - 6 years, 8 months ago

This does not prove that the limit indeed exists.

Abhishek Sinha - 6 years, 8 months ago

Log in to reply

The question just wanted the answer, and I don't know anything about real analysis or topology, or whatever branch of math you used. So, I just used some algebra to solve it. But, that is a valid point, since it may not have existed...

L N - 6 years, 8 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...