Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Power series expansion

1 view
Skip to first unread message

HardySpicer

unread,
Mar 2, 2011, 1:19:39 AM3/2/11
to
I have a polynomial

d(x)= 1+d1x+d2x^2 +...+ dnx^n

and I wish to express

p0+p1x+p2x^2+...+ pmx^m. = 1/d(x)

If I multiply out and compare coefficients I can come up with a
formula but I am unsure whether the limit on the summation is
different when m>n and m<n?


H.

Ray Vickson

unread,
Mar 2, 2011, 4:37:47 AM3/2/11
to

The reciprocal of a polynomial is an *infinite* series, so m =
infinity. For example, if |x| < 1 then 1/(1-x) = 1 + x + x^2 + x^3
+ .... with no end to the terms.

R.G. Vickson

Rob Johnson

unread,
Mar 2, 2011, 9:08:09 AM3/2/11
to
In article <3dc23e7d-f2e6-44ea...@j9g2000prj.googlegroups.com>,

Indeed. Say for example that we have two polynomials that were
reciprocals,

d(x) = 1 + d_1 x + d_2 x^2 + ... + d_n x^n

p(x) = 1 + p_1 x + p_2 x^2 + ... + p_m x^m

where d_n != 0 and p_m != 0. Then we have that the coefficient of
the x^{n+m} term of the product would be d_n p_m != 0. However, the
coefficient of any term, except the x^0 term, must be 0. Thus, we
reach a contradiction unless n+m = 0.

Thus, the only polynomial d(x) which can have a polynomial inverse
is d(x) = 1 where p(x) = 1.

Rob Johnson <r...@trash.whim.org>
take out the trash before replying
to view any ASCII art, display article in a monospaced font

0 new messages