Cosine and Something Else

Geometry Level 5

Let n n be a positive nonzero integer and

k = 4 n , k=4n,
a = x + x 2 k 2 , a=x+\sqrt { { x }^{ 2 }-{ k }^{ 2 } },
b = x x 2 k 2 , b=x-\sqrt { { x }^{ 2 }-{ k }^{ 2 } },
f ( x ) = 1 2 1 k k ( a k + b k ) . f\left( x \right) =\dfrac { 1 }{ 2 } \dfrac { 1 }{ { k }^{\displaystyle k } } \left( { a }^{ \displaystyle k }+{ b }^{\displaystyle k } \right).

For what integer n < 1000 n<1000 does f ( x ) f\left( x \right) intersect cos ( x ) \cos\left( x \right) the most times between 4 π < x < 4 π -4\pi <x<4\pi ?

You may want to use a graphing calculator.


The answer is 3.

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.

3 solutions

Michael Mendrin
Oct 4, 2014

Given that k = 4 n k=4n for postive integer n n , we have

a = x + x 2 k 2 = x + i k 2 x 2 = k ( x k + i 1 ( x k ) 2 ) a=x+\sqrt { { x }^{ 2 }-{ k }^{ 2 } } =x+i\sqrt { { k }^{ 2 }-{ x }^{ 2 } } =k\left( \dfrac { x }{ k } +i\sqrt { 1-{ \left( \dfrac { x }{ k } \right) }^{ 2 } } \right)

b = x x 2 k 2 = x i k 2 x 2 = k ( x k i 1 ( x k ) 2 ) b=x-\sqrt { { x }^{ 2 }-{ k }^{ 2 } } =x-i\sqrt { { k }^{ 2 }-{ x }^{ 2 } } =k\left( \dfrac { x }{ k } -i\sqrt { 1-{ \left( \dfrac { x }{ k } \right) }^{ 2 } } \right)

We then take advantage of this trigonometric identity

i A r c S i n ( x ) = L o g ( i x + 1 x 2 ) iArcSin\left( x \right) =Log\left( ix+\sqrt { 1-{ x }^{ 2 } } \right)

to rewrite a , b a,b as

a = k i e i ϕ a=ki{ e }^{\displaystyle -i\phi }

b = k i e i ϕ b=-ki{ e }^{\displaystyle i\phi }

where

ϕ = A r c S i n ( x k ) \phi =ArcSin\left( \dfrac { x }{ k } \right)

Given that a b = k 2 ab=k^{ 2 } and plugging in the expressions for a , b a,b , we end up with

1 2 1 k k ( a k + b k ) = 1 2 ( ( a b ) 1 2 k + ( b a ) 1 2 k ) = 1 2 ( e i k ϕ + e i k ϕ ) \dfrac { 1 }{ 2 } \dfrac { 1 }{ { k }^{ k } } \left( { a }^{ k }+{ b }^{ k } \right) =\dfrac { 1 }{ 2 } \left( { \left( \dfrac { a }{ b } \right) }^{ \frac { 1 }{ 2 } k }+{ \left( \dfrac { b }{ a } \right) }^{ \frac { 1 }{ 2 } k } \right) =\dfrac { 1 }{ 2 } \left( { e }^{\displaystyle -ik\phi }+{ e }^{ \displaystyle ik\phi } \right)

which, as k k\rightarrow \infty , becomes the exponential expression for C o s ( x ) Cos\left( x \right) . Keep in mind that k = 4 n k=4n , which eliminates pesky sign problems. Thus, as n n increases, it approaches the Cosine function, and from n = 4 n=4 on, it never crosses the Cosine function more than 4 4 times between 0 < x < 4 π 0<x<4\pi . For n = 3 n=3 , it crosses the Cosine function 5 5 times between 0 < x < 4 π 0<x<4\pi . Both functions are symmetric functions of x x .

Here's the graphic for n = 6 n=6 , the green vertical line is where x = 4 π x=4\pi

But values of x less than k (in magnitude) do not lie in the domain of the function.

Aalap Shah - 6 years, 8 months ago

Log in to reply

For k=4n, where n is an integer, complex quantities vanish, and we're left with real numbers, so x < k are in the domain. f(x) always produces a finite polynomial with rational coefficients.

Michael Mendrin - 6 years, 8 months ago

Hi Michael , Can you please look into my post???

Siddhartha Devapujula - 6 years, 8 months ago

Log in to reply

I am unable to understand the mistake in mine..

Siddhartha Devapujula - 6 years, 8 months ago

I am able to prove that both the curves never meet and it also seems to be correct. Don't really know where I have did any mistake in the proof. Kindly go through it and comment on it....

I am new to this site and didn't get accquainted with formatting stuff.... Sorry for that

Given k= 4n

and clearly x^2 > k^2

LHS = (1/2)*((a/k)^k + (b/k)^k)

(a/k)^k * (b/k)^k = 1

and both are positive numbers as K is an even number

So by AM>=GM

we have LHS >=1

RHS = cos(x)

RHS<=1

So this condition satisfies when LHS = RHS =1

LHS = 1 => a=b=x=+/- k

so we have X = +/- K when both the curves meet

and cos(x) = 1

=> x = 2mPi or -(2m+1)Pi

Case 1 : X= 2mPi

=> X=K as K cannot be negative

=> 2mPi = K

=> Pi = K/(2m)

But we know Pi is an irrational number and cannot be expressed as m/n

Similarly for Case also

Case 2: X= -(2m+1)Pi => X=-K

=>K = (2m+1)Pi

Here also Pi condraticts the rules of irrationality

So we can say that both the curves never meet........

No, it's not true that x^2 > k^2. As I pointed out in my reply to Aalap, when expanded, f(x) becomes a polynomial of finite order and with rational coefficients, with a real value even when x^2 < k^2. Your proof depends on that x^2 > k^2 always has to be true.

Michael Mendrin - 6 years, 8 months ago

Log in to reply

Great!!! I am unnecessarily assuming 'a' and 'b' have to be rational..... Thanks for the reply

Siddhartha Devapujula - 6 years, 8 months ago
Aditya Sankalp
Oct 9, 2014

first check that needs to be done is to see if f(x) is real for all values of x & k. We can see that for that

x 2 k 2 0 \sqrt { { x }^{ 2 }-{ k }^{ 2 } } \ge \quad 0

Checking this condition we get

x 2 k 2 0 x 2 k 2 0 ( x k ) ( x + k ) 0 x k x x v a r i e s f r o m 4 π t o + 4 π h e n c e 4 π k 4 π n o w k = 4 n s o 4 π 4 n 4 π g i v e s π n π S i n c e n i s a n i n t e g e r < 1000 s o t h e o n l y p o s s i b l e v a l u e s o f n c a n b e 1 , 2 , 3 f o r n = 1 f ( x ) = ( 1 2 1 4 4 ) . ( ( x + x 2 16 ) 4 + ( x x 2 16 ) 4 a g a i n w e c a n s e e t h a t x 4 o r x 4 f o r x = 4 f ( x ) = 1 f o r x > 4 o r x < 4 f ( x ) > 1 t h u s f ( x ) c a n i n t e r s e c t c o s ( x ) a t o n l y x = 4 = 1.272 π c o s ( x ) l i e s b e t w e e n 1 t o 0 f o r π t o 1.5 π . H e n c e n = 1 i s n o t a s o l u t i o n S i m i l a r l y n = 2 & n = 3 i s c h e c k e d . f o r n = 2 x = 2.54 π ; c o s ( x ) l i e s b e t w e e n 0 & 1 f o r 2.5 π t o 3 π f o r n = 3 x = 3.82 π ; c o s ( x ) l i e s b e t w e e n 0 & 1 f o r 3.5 π t o 4 π T h u s t h e r e i s o n l y o n e s o l u t i o n n = 3 \sqrt { { x }^{ 2 }-{ k }^{ 2 } } \ge \quad 0\quad \\ { x }^{ 2 }-{ k }^{ 2 }\ge 0\\ (x-k)(x+k)\ge 0\\ -x\le k\le x\\ x\quad varies\quad from\quad -4\pi \quad to\quad +4\pi \\ hence\quad \\ -4\pi \le k\le 4\pi \\ \\ now\quad k\quad =\quad 4n\\ so\quad -4\pi \le 4n\le 4\pi \\ \\ gives\quad -\pi \le n\le \pi \\ \\ Since\quad n\quad is\quad an\quad integer\quad <1000\quad so\quad the\quad only\quad possible\quad values\quad of\quad n\quad can\quad be\quad 1,2,3\\ \\ \\ for\quad n\quad =\quad 1\quad \\ \\ f(x)\quad =\quad (\frac { 1 }{ 2 } \frac { 1 }{ { 4 }^{ 4 } } ).((x+{ \sqrt { { x }^{ 2 }-16 } }){ ^{ 4 } }+(x-{ \sqrt { { x }^{ 2 }-16 } }){ ^{ 4 } }\\ again\quad we\quad can\quad see\quad that\quad x\quad \ge 4\quad or\quad x\quad \le -4\\ for\quad x=4\quad f(x)\quad =\quad 1\\ for\quad x>4\quad or\quad x\quad <-4\quad f(x)\quad >1\\ thus\quad f(x)\quad can\quad intersect\quad cos(x)\quad at\quad only\quad x\quad =\quad 4\quad =\quad 1.272\pi \\ cos(x)\quad lies\quad between\quad -1\quad to\quad 0\quad for\quad \pi \quad to\quad 1.5\pi .\\ \\ Hence\quad n\quad =\quad 1\quad is\quad not\quad a\quad solution\\ \\ Similarly\quad n\quad =\quad 2\quad \& \quad n\quad =\quad 3\quad is\quad checked.\\ for\quad n\quad =\quad 2\quad x\quad =\quad 2.54\pi \quad ;\quad cos(x)\quad lies\quad between\quad 0\quad \& \quad -1\quad for\quad 2.5\pi \quad to\quad 3\pi \\ \\ for\quad n\quad =\quad 3\quad x\quad =\quad 3.82\pi \quad ;\quad cos(x)\quad lies\quad between\quad 0\quad \& \quad 1\quad for\quad 3.5\pi \quad to\quad 4\pi \\ \\ Thus\quad there\quad is\quad only\quad one\quad solution\quad n\quad =\quad 3

The solution has been found using elimination method. A detailed way wold be to plot the equations. However elimination method is quite useful to narrow down the possible solutions.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...