lo g 2 [ cos ( 2 0 1 5 2 π ) cos ( 2 0 1 5 4 π ) cos ( 2 0 1 5 6 π ) ⋯ cos ( 2 0 1 5 4 0 2 8 π ) ] = ?
Bonus : Generalize k = 0 ∏ 2 n cos ( 2 n + 1 2 k π ) in terms of n with a proper derivation without using methods like Induction.
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.
Or you can show that the expression inside the log is product of cosine of 1,2,3,... 2014 times pi/2015, then apply Chebyshev polynomials.
Log in to reply
We all have our favourites, Comrade Pi. As you know, I love my roots of unity... (and by the way, the "product of cosine of 1,2,3,... 2014 times pi/2015" is going to be negative)
Let z = e i θ = x + i y . Then:
cos ( ( 2 n + 1 ) θ ) = R e ( z 2 n + 1 ) = k = 0 ∑ n ( 2 k 2 n + 1 ) x 2 ( n − k ) + 1 ( i y ) 2 k = k = 0 ∑ n ( − 1 ) k ( 2 k 2 n + 1 ) x 2 ( n − k ) + 1 ( 1 − x 2 ) k
which we denote as f ( x ) . It's easy to see that:
f ( x ) = a 2 n + 1 x 2 n + 1 + … + a 1 x 1 . . . ( 1 )
where
a 1 = ( − 1 ) n ( 2 n 2 n + 1 ) = ( − 1 ) n ( 2 n + 1 ) . . . ( 2 )
On the other hand, since cos ( ( 2 n + 1 ) θ ) = 2 1 ( z 2 n + 1 + z 2 n + 1 1 ) = 1 + 2 1 ( z 2 n + 1 + z 2 n + 1 1 − 2 )
= 1 + 2 z 2 n + 1 1 ( z 2 n + 1 − 1 ) 2 and
( z 2 n + 1 − 1 ) 2 = k = 0 ∏ 2 n ( z − e 2 n + 1 2 k π i ) 2 = k = 0 ∏ 2 n ( z − e 2 n + 1 2 k π i ) ( z − e 2 n + 1 − 2 k π i )
= k = 0 ∏ 2 n ( z 2 + 1 − 2 z cos ( 2 n + 1 2 k π ) )
We also have:
f ( x ) = 1 + 2 1 k = 0 ∏ 2 n ( z + z 1 − 2 cos ( 2 n + 1 2 k π ) )
= 1 + 2 2 n k = 0 ∏ 2 n ( x − 2 cos ( 2 n + 1 2 k π ) ) . . . ( 3 )
Now we compare the expressions ( 1 ) and ( 3 ) for f ( x ) . We see from ( 1 ) that the constant term in ( 3 ) is zero. Therefore:
2 2 n k = 0 ∏ 2 n cos ( 2 n + 1 2 k π ) = 1 ⟹ k = 0 ∏ 2 n cos ( 2 n + 1 2 k π ) = 2 − 2 n
Thus, in the problem, we are asked the value of:
lo g 2 [ k = 0 ∏ 2 0 1 4 cos ( 2 0 1 5 2 k π ) ] = lo g 2 ( 2 − 2 0 1 4 ) = − 2 0 1 4
Lemma 1 : r = 1 ∏ n cos 2 n + 1 r π = 2 n 1
Proof : Let, C = r = 1 ∏ n cos ( 2 n + 1 r π )
and
S = r = 1 ∏ n sin ( 2 n + 1 r π )
Now,
C ⋅ S = ( sin 2 n + 1 π ⋅ cos 2 n + 1 π ) ⋅ ( sin 2 n + 1 2 π ⋅ cos 2 n + 1 2 π ) ⋅ … ⋅ ( sin 2 n + 1 n π ⋅ cos 2 n + 1 n π )
⟹ C ⋅ S = 2 n 1 ( 2 sin 2 n + 1 π ⋅ cos 2 n + 1 π ) ⋅ ( 2 sin 2 n + 1 2 π ⋅ cos 2 n + 1 2 π ) ⋅ … ⋅ ( 2 sin 2 n + 1 n π ⋅ cos 2 n + 1 n π )
⟹ C ⋅ S = 2 n 1 sin 2 n + 1 2 π ⋅ sin 2 n + 1 4 π ⋅ … ⋅ sin 2 n + 1 2 n π
{ ∵ sin ( 2 x ) = 2 sin ( x ) cos ( x ) }
⟹ C ⋅ S = 2 n 1 sin 2 n + 1 π ⋅ sin 2 n + 1 2 π ⋅ … ⋅ sin 2 n + 1 n π { ∵ sin ( π − x ) = sin ( x ) }
⟹ C ⋅ S = 2 n 1 ⋅ S
since S = 0 ,
∴ C = 2 n 1
Lemma 2 : r = 1 ∏ 2 n cos 2 n + 1 2 r π = ( r = 1 ∏ n cos 2 n + 1 r π ) 2
Proof : r = 1 ∏ 2 n cos 2 n + 1 2 r π = ( r = 1 ∏ n cos 2 n + 1 2 r π ) ⋅ ( r = n + 1 ∏ 2 n cos 2 n + 1 2 r π )
= ( − 1 ) n ( r = 1 ∏ n cos 2 n + 1 2 r π ) ⋅ ( r = 1 ∏ n cos 2 n + 1 ( 2 r − 1 ) π ) { ∵ cos ( π + x ) = − cos ( x ) }
= ( − 1 ) n r = 1 ∏ 2 n cos 2 n + 1 r π
= ( − 1 ) n ( r = 1 ∏ n cos 2 n + 1 r π ) ⋅ ( r = n + 1 ∏ 2 n cos 2 n + 1 r π )
= ( − 1 ) n ( r = 1 ∏ n cos 2 n + 1 r π ) ⋅ ( r = 1 ∏ n cos 2 n + 1 ( n + r ) π )
= ( − 1 ) n ( r = 1 ∏ n cos 2 n + 1 r π ) ⋅ ( r = 1 ∏ n cos 2 n + 1 ( 2 n + 1 − r ) π ) ( ∵ r = 1 ∏ n f ( r ) = r = 1 ∏ n f ( n + 1 − r ) )
= ( r = 1 ∏ n cos 2 n + 1 r π ) ⋅ ( r = 1 ∏ n cos 2 n + 1 r π ) = ( r = 1 ∏ n cos 2 n + 1 r π ) 2 { ∵ cos ( π − x ) = cos x }
Combining the above two Lemmas, we have,
r = 0 ∏ 2 n cos 2 n + 1 2 r π = 2 2 n 1
∴ lo g 2 [ r = 0 ∏ 2 0 1 4 cos ( 2 0 1 5 2 r π ) ] = lo g 2 ( 2 − 2 0 1 4 ) = − 2 0 1 4
Let w = e 4 π i / n where n is odd, and let S = k = 1 ∏ n − 1 cos ( n 2 π k )
Since cos ( θ ) = cos ( 2 π − θ ) , we know that S is always positive.
Now, with the identity ∣ 1 + w k ∣ = 2 ∣ ∣ ∣ ∣ cos ( n 2 π k ) ∣ ∣ ∣ ∣ that becomes to:
S = k = 1 ∏ n − 1 2 ∣ 1 + w k ∣ = 2 n − 1 ∣ 1 + w ∣ ∣ 1 + w 2 ∣ ⋯ ∣ 1 + w ( n − 1 ) ∣
But also P ( x ) = x − 1 x n − 1 = ( x − w ) ( x − w 2 ) ⋯ ( x − w n − 1 ) , that implies that x n − 1 + x n − 2 + ⋯ + x + 1 = ( x − w ) ( x − w 2 ) ⋯ ( x − w n − 1 ) . If we let x = − 1 we obtain: 1 = ( − 1 − w ) ( − 1 − w 2 ) ⋯ ( − 1 − w n − 1 ) , or 1 = ( 1 + w ) ( 1 + w 2 ) ⋯ ( 1 + w n − 1 ) .
Finally, apply absolute value to both sides:
1 = ∣ 1 + w ∣ ∣ 1 + w 2 ∣ ⋯ ∣ 1 + w n − 1 ∣
So, the product is: S = 2 n − 1 1
In this case n = 2 0 1 5 , so lo g 2 ( 2 − 2 0 1 4 ) = − 2 0 1 4 .
Are you saying that ∏ k = 1 n − 1 cos ( n k π ) = ∏ k = 1 n − 1 cos ( n 2 k π ) for all odd n ?
Log in to reply
No, I missed a sign. I've updated my solution.
Log in to reply
Yes, that makes good sense now (upvoted)! Both of our solutions are based on the fact that e 4 π i / n is a primitive n th root of unity.
Log in to reply
@Otto Bretscher – Thanks :D Your solution is also good, you are the master of the roots of unity. Also upvoted.
Log in to reply
@Alan Enrique Ontiveros Salazar – Far from being a master in the field, I'm just an enthusiastic amateur ;)
Problem Loading...
Note Loading...
Set Loading...
Let m = 2 n + 1 be an odd positive integer, and consider w = e 2 π i / m .
Now z m − 1 = ( z − 1 ) ∏ k = 1 m − 1 ( z − w k ) . Substituting z = − 1 and simplifying a bit, we find that ∏ k = 1 m − 1 ( 1 + w k ) = 1 .
Now ∏ k = 1 m − 1 cos ( m 2 k π ) = ∏ k = 1 m − 1 2 1 ( w k + w − k ) = 2 m − 1 1 ∏ k = 1 m − 1 ( w k + w − k ) = 2 m − 1 1 ∏ k = 1 m − 1 ( w 2 k + 1 ) = 2 m − 1 1 ∏ k = 1 m − 1 ( w k + 1 ) = 2 m − 1 1 .
The penultimate step follows from the fact that w 2 is a primitive m th root of unity. The third step from the back follows from the fact that the product of the m th roots of unity is 1, by Viete.