<< Chapter < Page Chapter >> Page >
We use Taylor's Remainder Theorem to derive a generalization of the Binomial Theorem to nonintegral exponents. First we must generalize the definition of binomial coefficient.

We use Taylor's Remainder Theorem to derive a generalization of the Binomial Theorem to nonintegral exponents. First we must generalize the definition of binomial coefficient.

Let α be a complex number, and let k be a nonnegative integer. We define the general binomial coefficient α k by

α k = α ( α - 1 ) ... ( α - k + 1 ) k ! .

If α is itself a positive integer and k α , then α k agrees with the earlier definition of the binomial coefficient,and α k = 0 when k > α . However, if α is not an integer, but just an arbitrary complex number, then every α k 0 .

Estimates for the size of binomial coefficients. Let α be a fixed complex number.

  1. Show that
    | α k | j = 1 k ( 1 + | α | j )
    for all nonnegative integers k . HINT: Note that
    | α k | | α | ( | a l p h a | + 1 ) ( | a l p h a | + 2 ) ... ( | α | + k - 1 ) k ! .
  2. Use part (a) to prove that there exists a constant C such that
    | α k | C 2 k
    for all nonnegative integers k . HINT: Note that ( 1 + | α | / j ) < 2 for all j > | α | .
  3. Show in fact that for each ϵ > 0 there exists a constant C ϵ such that
    | α k | C ϵ ( 1 + ϵ ) k
    for all nonnegative integers k .
  4. Let h ( t ) be the power series function given by h ( t ) = k = 0 α k t k . Use the ratio test to show that the radius of convergence for h equals 1.

REMARK The general Binomial Theorem, if there is one, should be something like the following:

( x + y ) α = k = 0 α k x α - k y k .

The problem is to determine when this infinite series converges, i.e., for what values of the three variables x , y , and α does it converge. It certainly is correct if x = 0 , so we may as well assume that x 0 , in which case we are considering the validity of the formula

( x + y ) α = x α ( 1 + t ) α = x α k = 0 α k t k ,

where t = y / x . Therefore, it will suffice to determine for what values of t and α does the infinite series

k = 0 α k t k

equal

( 1 + t ) α .

The answer is that, for n arbitrary complex number α , this series converges to the correct value for all t ( - 1 , 1 ) . (Of course, t must be larger than - 1 for the expression ( 1 + t ) α even to be defined.) However, the next theorem only establishes this equality for t 's in the subinterval ( - 1 / 2 , 1 / 2 ) . As mentioned earlier, its proof is based on Taylor's Remainder Theorem. We must postpone the complete proof to [link] , where we will have a better version of Taylor's Theorem.

Let α = a + b i be a fixed complex number. Then

( 1 + t ) α = k = 0 α k t k

for all t ( - 1 / 2 , 1 / 2 ) .

Of course, this theorem is true if α is a nonnegative integer, for it is then just the original Binomial Theorem,and in fact in that case it holds for every complex number t . For a general complex number α , we have only defined x α for positive x 's, so that ( 1 + t ) α is not even defined for t < - 1 .

Now, for a general α = a + b i , consider the function g : ( - 1 / 2 , 1 / 2 ) C defined by g ( t ) = ( 1 + t ) α . Observe that the n th derivative of g is given by

g ( n ) ( t ) = α ( α - 1 ) ... ( α - n + 1 ) ( 1 + t ) n - α .

Then g C ( ( - 1 / 2 , 1 / 2 ) ) . (Of course, g is actually in C ( - 1 , 1 ) , but the present theorem is only concerned with t 's in ( - 1 / 2 , 1 / 2 ) . )

For each nonnegative integer k define

a k = g ( k ) ( 0 ) / k ! = α ( α - 1 ) ... ( α - k + 1 ) k ! = α k ,

and set h equal to the power series function given by h ( t ) = k = 0 a k t k . According to part (d) of the preceding exercise, the radius of convergence for the power series a k t k is 1. The aim of this theorem is toshow that g ( t ) = h ( t ) for all - 1 / 2 < t < 1 / 2 . In other words, we wish to show that g agrees with this power series function at least on the interval ( - 1 / 2 , 1 / 2 ) . It will suffice to show that the sequence { S n } of partial sums of the power series function h converges to the function g , at least on ( - 1 / 2 , 1 / 2 ) . We note also that the n th partial sum of this power series is just the n th Taylor polynomial T g n for g .

S n ( t ) = k = 0 n α k t k = k = 0 n g ( k ) ( 0 ) k ! t k .

Now, fix a t strictly between - 1 / 2 and 1 / 2 , and let r < 1 be as in part (c) of [link] . That is, | t / ( 1 + y ) | < r for every y between 0 and t . (This is an important inequality for our proof, and this is one place where the hypothesis that t ( - 1 / 2 , 1 / 2 ) is necessary.) Note also that, for any y ( - 1 / 2 , 1 / 2 ) , we have | ( 1 + y ) α | = ( 1 + y ) a , and this is trapped between ( 1 / 2 ) a and ( 3 / 2 ) a . Hence, there exists a number M such that | ( 1 + y ) α | M for all y ( - 1 / 2 , 1 / 2 ) .

Next, choose an ϵ > 0 for which β = ( 1 + ϵ ) r < 1 . We let C ϵ be a constant satisfying the inequality in Part (c) of [link] . So, using Taylor's Remainder Theorem, we have thatthere exists a y between 0 and t for which

| g ( t ) - k = 0 n a k t k | = | g ( t ) - ( T ( g , 0 ) n ( t ) | = | g ( n + 1 ) ( y ) ( n + 1 ) ! t n + 1 | = | α ( α - 1 ) ... ( α - n ) ( n + 1 ) ! ( 1 + y ) n + 1 - α t n + 1 | | α n + 1 | | ( 1 + y ) α | | t 1 + y | n + 1 C ϵ ( 1 + ϵ ) n + 1 M | t 1 + y | n + 1 C ϵ ( 1 + ϵ ) n + 1 M r n + 1 C ϵ M β n + 1 , .

Taking the limit as n tends to , and recalling that β < 1 , shows that g ( t ) = h ( t ) for all - 1 / 2 < t < 1 / 2 , which completes the proof.

Get Jobilize Job Search Mobile App in your pocket Now!

Get it on Google Play Download on the App Store Now




Source:  OpenStax, Analysis of functions of a single variable. OpenStax CNX. Dec 11, 2010 Download for free at http://cnx.org/content/col11249/1.1
Google Play and the Google Play logo are trademarks of Google Inc.

Notification Switch

Would you like to follow the 'Analysis of functions of a single variable' conversation and receive update notifications?

Ask