Limit of a cosine's function iteration

jwpaine

Full Member
Joined
Mar 10, 2007
Messages
723
I need help wraping my head around how to do this, and /why/ it works out the way it does.

The purpose of this exercise is to illustrate how a transcendental equation can be solved using limits. Let f(x) = cosx for x E R We intend to solve the equation f(x) = x. Denote the solution by \(\displaystyle \x\). Then \(\displaystyle \x\) is the unique real number such that \(\displaystyle f(\x) = \x\). To find \(\displaystyle \x\), we start with an initial quess \(\displaystyle x_0\). The choice of \(\displaystyle x_0\) isn't too important. Any number in the interval [-10..10] should work, but for simplicity you take \(\displaystyle x_0 = 0\). Now calculate

\(\displaystyle \L s_1 = f(x_0), \,\,\,x_2 = f(x_1) = f(f(x_0)),\,\,\, x_3 = f(x_2) = f(f(f(x_0))), ...\)

\(\displaystyle \L x_n = f(x_{n-1}) = f(...f(f(x_0))...)\)

You can do this by repeatedly pushing the cosine button on your calculator. Be sure to use radian mode. After 20 cosine evaluations (n = 20) you should have a number \(\displaystyle x_{20}\) whose first two or three decimal digits agree with those of the actual solution \(\displaystyle \x\) Infact: \(\displaystyle \L \x = \lim_{x\to\infty}x_n\)



Give the first 6 decimal digits of \(\displaystyle \x\)



First of all - I really don't understand /why/ this infinite function composition approaches a specific positive value: If I graph f(x), f(f(x), f(f(f(x))), .... I see the cosine function stretching out into /almost/ a horizontal line at around 0.7389

EDIT: And how would I show this without just doing cos(cos(cos(x))).. on my calculator? I know this is not the place for tutoring, but I hope someone can help explain this to me.
 
Hey JW.

If you set \(\displaystyle \L\\x=cos(x)\) and solve for x, you'll find x=0.739085.......

You can use Newton's method to find the convergence of x-cos(x)=0

\(\displaystyle \L\\x_{n+1}=x_{n}-\frac{f(x)}{f'(x)}\)

You will see it converges to what you found, 0.739085.............

It converges much faster with Newton's method than with the infinite compositions, but the principle is the same. You're honing in on a point.

Look at the graph of y=cos(x) and y=x. They intersect at x=0.739085
 
Here's the graph:

cosxwn9.jpg
 
Top