A curious identity

Algebra Level 5

Let w = e π i / 2015 . w=e^{\pi{i}/2015}. Without resorting to trigonometry, find k = 1 2014 ( w k w k ) \prod_{k=1}^{2014}(w^k-w^{-k})


The answer is -2015.

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.

1 solution

Pradeep Maurya
May 6, 2015

k = 1 2014 ( w k w k ) = k = 1 2014 ( w 2 k 1 ) w k = 1 2014 k = k = 1 2014 ( e 2 k π i 2015 1 ) e 1007 π i \prod_{k = 1}^{2014} \left(w^{k}-w^{-k}\right) = \displaystyle \frac{\prod_{k = 1}^{2014} \left(w^{2k}-1\right)}{w^{\sum_{k = 1}^{2014} k}} = \frac{\prod_{k = 1}^{2014} \left(e^{\frac{2k\pi i}{2015}}-1\right)}{ e^{1007 \pi i}}

Now consider the set { e 2 k π i 2015 : k = 0 , 1 , 2 , . . . . , 2014 } \{e^{\frac{2k\pi i}{2015}}: k = 0,1,2,....,2014\} is the set of all solutions of z 2015 = 1 z^{2015} = 1 Hence

z 2015 1 = ( z 1 ) k = 1 2014 ( z e 2 k π i 2015 ) z^{2015}-1 = (z-1) \prod_{k=1}^{2014} \left(z- e^{\frac{2k\pi i}{2015}}\right)

1 + z + z 2 + . . . . z 2014 = k = 1 2014 ( z e 2 k π i 2015 ) 1+z+z^2+....z^{2014} = \prod_{k = 1}^{2014}\left(z - e^{\frac{2k\pi i}{2015}} \right)

Putting z = 1 z = 1 both sides gives k = 1 2014 ( e 2 k π i 2015 1 ) = 2015 \prod_{k=1}^{2014} \left(e^{\frac{2k\pi i}{2015}}-1\right) = 2015

Hence answer is 2015 \boxed{-2015}

And from there, we can actually derive -

k = 1 n 1 s i n ( k π n ) = n 2 n 1 \displaystyle \prod_{k = 1}^{n-1}{sin(\frac{k\pi}{n})} = \frac{n}{{2}^{n-1}}

*You must have gotten how.

Kartik Sharma - 6 years, 1 month ago

Log in to reply

Yes, exactly... that's why I posted this problem, to outline a simple proof for this trig identity. Let me do the same thing for the cos identity.

Otto Bretscher - 6 years, 1 month ago

I believe it is a product, not a sum.

Jake Lai - 6 years, 1 month ago

Log in to reply

Oops! A typo! Thanks for pointing it out!

Kartik Sharma - 6 years, 1 month ago

Yes indeed...In the solution I just took care of the part "without resorting to trigonometry".

Pradeep Maurya - 6 years, 1 month ago

Thank you for posting this clear and elegant proof! That's exactly what I had in mind...

Otto Bretscher - 6 years, 1 month ago

Log in to reply

No idea why the answer eluded me for so long! I was searching for some explanation by Newton's identities, Vieta's formulas, even Euler's pentagonal number theorem (before I realised that was an infinite product).

Boiled down to polynomial interpolation in the end; such a simple but beautiful problem.

Jake Lai - 6 years, 1 month ago

You cannot divide by ( z 1 z-1 ) and then substitute z = 1 z=1 !

Kenny Lau - 5 years, 8 months ago

Log in to reply

Sure we can...we are dealing with a polynomial.

Otto Bretscher - 5 years, 8 months ago

In the second last step of the solution you got the answer 2015. Then how did you conclude that the answer to the question is -2015 ???

Mathews Boban - 6 years, 1 month ago

Log in to reply

He has a factor e 1007 π i = 1 e^{1007\pi{i}}=-1 in the first equation.

Otto Bretscher - 6 years, 1 month ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...