Hadamard matrix: Difference between revisions

From formulasearchengine
Jump to navigation Jump to search
en>Rswarbrick
→‎Properties: Explain where the transpose/inverse result comes from.
en>Wavelength
Line 1: Line 1:
In [[mathematics]], a '''telescoping series''' is a [[series (mathematics)|series]] whose partial sums eventually only have a fixed number of terms after cancellation.<ref>[[Tom M. Apostol]], ''Calculus, Volume 1,'' Blaisdell Publishing Company, 1962, pages&nbsp;422&ndash;3</ref><ref>Brian S. Thomson and Andrew M. Bruckner, ''Elementary Real Analysis, Second Edition'', CreateSpace, 2008, page 85</ref> Such a technique is also known as the '''method of differences'''.
I'm Randall (23) from Queenzieburn, Great Britain. <br>I'm learning Korean literature at a local high school and I'm just about to graduate.<br>I have a part time job in a post office.<br><br>Feel free to visit my webpage; [http://zhenyoukang.com/plus/guestbook.php Fifa 15 Coin Generator]
 
For example, the series
 
:<math>\sum_{n=1}^\infty\frac{1}{n(n+1)}</math>
 
simplifies as
 
:<math>\begin{align}
\sum_{n=1}^\infty \frac{1}{n(n+1)} & {} = \sum_{n=1}^\infty \left( \frac{1}{n} - \frac{1}{n+1} \right) \\
& {} = \lim_{N\to\infty} \sum_{n=1}^N \left( \frac{1}{n} - \frac{1}{n+1} \right) \\
& {} = \lim_{N\to\infty} \left \lbrack {\left(1 - \frac{1}{2}\right) + \left(\frac{1}{2} - \frac{1}{3}\right) + \cdots + \left(\frac{1}{N} - \frac{1}{N+1}\right) } \right \rbrack  \\
& {} = \lim_{N\to\infty} \left \lbrack {  1 + \left(- \frac{1}{2} + \frac{1}{2}\right) + \left( - \frac{1}{3} + \frac{1}{3}\right) + \cdots + \left( - \frac{1}{N} + \frac{1}{N}\right) - \frac{1}{N+1} } \right \rbrack = 1.
\end{align}</math>
 
==In general==
Let <math>a_n</math> be a sequence of numbers. Then,
 
:<math>\sum_{n=1}^N \left(a_n - a_{n-1}\right) =  a_N - a_{0},</math>
 
and, if <math>a_n \rightarrow 0</math>
 
:<math>\sum_{n=1}^\infty \left(a_n - a_{n-1}\right) =  - a_{0}.</math>
 
==A pitfall==
 
Although telescoping can be a useful technique, there are pitfalls to watch out for:
 
:<math>0 = \sum_{n=1}^\infty 0 = \sum_{n=1}^\infty (1-1) = 1 + \sum_{n=1}^\infty (-1 + 1) = 1\,</math>
 
is not correct because this regrouping of terms is invalid unless the individual terms [[Convergent sequence|converge]] to 0; see [[Grandi's series]].  The way to avoid this error is to find the sum of the first ''N'' terms first and ''then'' take the limit as ''N'' approaches infinity:
 
:<math>
\begin{align}
\sum_{n=1}^N \frac{1}{n(n+1)} & {} = \sum_{n=1}^N \left( \frac{1}{n} - \frac{1}{n+1} \right) \\
& {} = \left(1 - \frac{1}{2}\right) + \left(\frac{1}{2} - \frac{1}{3}\right) + \cdots + \left(\frac{1}{N} -\frac{1}{N+1}\right) \\
& {} =  1 + \left(- \frac{1}{2} + \frac{1}{2}\right)
+ \left( - \frac{1}{3} + \frac{1}{3}\right) + \cdots
+ \left(-\frac{1}{N} + \frac{1}{N}\right) - \frac{1}{N+1} \\
& {} = 1 - \frac{1}{N+1}\to 1\ \mathrm{as}\ N\to\infty.
\end{align}
</math>
 
==More examples==
 
* Many [[trigonometric function]]s also admit representation as a difference, which allows telescopic cancelling between the consecutive terms.
 
::<math>
\begin{align}
\sum_{n=1}^N \sin\left(n\right) & {} = \sum_{n=1}^N \frac{1}{2} \csc\left(\frac{1}{2}\right) \left(2\sin\left(\frac{1}{2}\right)\sin\left(n\right)\right) \\
& {} =\frac{1}{2} \csc\left(\frac{1}{2}\right) \sum_{n=1}^N \left(\cos\left(\frac{2n-1}{2}\right) -\cos\left(\frac{2n+1}{2}\right)\right) \\
& {} =\frac{1}{2} \csc\left(\frac{1}{2}\right) \left(\cos\left(\frac{1}{2}\right) -\cos\left(\frac{2N+1}{2}\right)\right).
\end{align}
</math>
 
* Some sums of the form
 
::<math>\sum_{n=1}^N {f(n) \over g(n)},</math>
 
:where ''f'' and ''g'' are [[polynomial function]]s whose quotient may be broken up into [[partial fraction]]s, will fail to admit [[summation]] by this method.  In particular, we have
 
::<math>
\begin{align}
\sum^\infty_{n=0}\frac{2n+3}{(n+1)(n+2)} & {} =\sum^\infty_{n=0}\left(\frac{1}{n+1}+\frac{1}{n+2}\right) \\
& {} = \left(\frac{1}{1} + \frac{1}{2}\right) + \left(\frac{1}{2} + \frac{1}{3}\right) + \left(\frac{1}{3} + \frac{1}{4}\right) + \cdots \\
& {} \cdots + \left(\frac{1}{n-1} + \frac{1}{n}\right) + \left(\frac{1}{n} + \frac{1}{n+1}\right) + \left(\frac{1}{n+1} + \frac{1}{n+2}\right) + \cdots \\
& {} =\infty.
\end{align}
</math>
 
:The problem is that the terms do not cancel.
 
* Let ''k'' be a positive integer.  Then
 
::<math>\sum^\infty_{n=1} {\frac{1}{n(n+k)}} = \frac{H_k}{k} </math>
 
:where ''H''<sub>''k''</sub> is the ''k''th [[harmonic number]].  All of the terms after 1/(''k''&nbsp;&minus;&nbsp;1) cancel.
 
== An application in probability theory ==
 
In [[probability theory]], a [[Poisson process]] is a stochastic process of which the simplest case involves "occurrences" at random times, the waiting time until the next occurrence having a [[memorylessness|memoryless]] [[exponential distribution]], and the number of "occurrences" in any time interval having a [[Poisson distribution]] whose expected value is proportional to the length of the time interval. Let ''X''<sub>''t''</sub> be the number of "occurrences" before time ''t'', and let ''T''<sub>''x''</sub> be the waiting time until the ''x''th "occurrence".  We seek the [[probability density function]] of the [[random variable]] ''T''<sub>''x''</sub>.  We use the [[probability mass function]] for the Poisson distribution, which tells us that
 
: <math> \Pr(X_t = x) = \frac{(\lambda t)^x e^{-\lambda t}}{x!}, </math>
 
where λ is the average number of occurrences in any time interval of length 1.  Observe that the event {''X''<sub>''t''</sub> ≥ x} is the same as the event {''T''<sub>''x''</sub> ≤ ''t''}, and thus they have the same probability. The density function we seek is therefore
 
: <math>
\begin{align}
f(t) & {} = \frac{d}{dt}\Pr(T_x \le t) = \frac{d}{dt}\Pr(X_t \ge x) = \frac{d}{dt}(1 - \Pr(X_t \le x-1)) \\  \\
& {} =  \frac{d}{dt}\left( 1 - \sum_{u=0}^{x-1} \Pr(X_t = u)\right)
= \frac{d}{dt}\left( 1 - \sum_{u=0}^{x-1} \frac{(\lambda t)^u e^{-\lambda t}}{u!}  \right) \\  \\
& {} = \lambda e^{-\lambda t} - e^{-\lambda t} \sum_{u=1}^{x-1} \left( \frac{\lambda^ut^{u-1}}{(u-1)!} - \frac{\lambda^{u+1} t^u}{u!} \right)
\end{align}
</math>
 
The sum telescopes, leaving
 
: <math> f(t) = \frac{\lambda^x t^{x-1} e^{-\lambda t}}{(x-1)!}. </math>
 
== Other applications ==
 
For other applications, see:
 
* [[Grandi's series]];
* [[Proof that the sum of the reciprocals of the primes diverges]], where one of the proofs uses a telescoping sum;
* [[Order statistic]], where a telescoping sum occurs in the derivation of a probability density function;
* [[Lefschetz fixed-point theorem]], where a telescoping sum arises in [[algebraic topology]];
* [[Homology theory]], again in algebraic topology;
* [[Eilenberg–Mazur swindle]], where a telescoping sum of knots occurs.
 
== Notes and references ==
{{reflist}}
 
{{DEFAULTSORT:Telescoping Series}}
[[Category:Mathematical series]]

Revision as of 03:30, 18 February 2014

I'm Randall (23) from Queenzieburn, Great Britain.
I'm learning Korean literature at a local high school and I'm just about to graduate.
I have a part time job in a post office.

Feel free to visit my webpage; Fifa 15 Coin Generator