2
$\begingroup$

Possible Duplicate:
How can we sum up $\sin$ and $\cos$ series when the angles are in arithmetic progression?

How is it possible to show for integer $m$:

$$\frac{1}{M}\sum_{k=1}^{M}\sin(m\cdot y_{k})=0$$

Thank you very much

Interval $[-\pi,\pi]$ split into $M$ equal intervals, with the mid point of interval is $y_{k}$

$\endgroup$
2
  • $\begingroup$ what is m, N? did you mean "for integer m"? if so, what about m=1? $\endgroup$
    – tomasz
    Jun 17, 2012 at 20:28
  • 2
    $\begingroup$ @Gerry: while abstractly I agree it is a duplicate, in this particular case the question is about summing an odd function at points symmetric about zero. $\endgroup$
    – Willie Wong
    Jun 20, 2012 at 8:01

2 Answers 2

Reset to default
4
$\begingroup$

The sine function is odd function, meaning that in general $\sin(-t)=-\sin(t)$.

The interval $[-\pi,\pi]$ is symmetrical about $0$. So if $x$ is one of the $x_k$, then $-x$ also is one of the $x_k$, and the values of $\sin(mx)$ and $\sin(-mx)$ add up to $0$. (If $N$ is odd, there isn't the perfect twinning, but the untwinned point is $0$, and $\sin(0)=0$.)

Exactly the same argument works for any odd function and any interval symmetric about the origin.

$\endgroup$
6
  • $\begingroup$ but how to proove this? $\endgroup$
    – nanme
    Jun 17, 2012 at 20:37
  • $\begingroup$ How to prove that $\sin x$ is an odd function? $\endgroup$
    – André Nicolas
    Jun 17, 2012 at 20:40
  • $\begingroup$ thank you very much but problem I try to apply concept to new problem but cosine function not symmmetric and so don't know how to show the statements truth.$$ \frac{1}{M}\sum_{j=1}^{M}\cos(mx_{j})=\begin{cases} 1, & \ m \equiv 0\pmod{M}\\ 0, & \text{else} \end{cases}$$ $\endgroup$
    – nanme
    Jun 17, 2012 at 20:52
  • $\begingroup$ @nanme: Different functions may require different techniques. My answer used a technique that works for all odd functions. The cosine problem above is not yet clear enough for me to give an answer. $\endgroup$
    – André Nicolas
    Jun 17, 2012 at 20:59
  • $\begingroup$ I understand. it follows from the same description as the original sin question. I try to find technique that is general and so can be used in many cases but am not sure if there is one that does exist. thank you again very much $\endgroup$
    – nanme
    Jun 17, 2012 at 21:09
1
$\begingroup$

When you sum over the $y_k$'s backwards, you get the both the same value and its negative:

$$\frac{1}{N}\sum_{k=1}^N \sin(m y_k) = \frac{1}{N} \sum_{k=1}^N \sin(m y_{N+1-k}) = \frac{1}{N} \sum_{k=1}^N \sin(-m y_k) = - \frac{1}{N} \sum_{k=1}^N \sin(m y_k)$$

so the sum is zero.

$\endgroup$
2
  • $\begingroup$ hi and thank you for response. however i don't seem to follow your logic thank you again $\endgroup$
    – nanme
    Jun 17, 2012 at 20:32
  • $\begingroup$ The $y_k$ are symmetric about 0 and $\sin(-x) = -\sin(x)$ for any $x$. $\endgroup$
    – Cocopuffs
    Jun 17, 2012 at 20:42

Not the answer you're looking for? Browse other questions tagged .