09.12.2012 Views

Concrete mathematics : a foundation for computer science

Concrete mathematics : a foundation for computer science

Concrete mathematics : a foundation for computer science

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

also has nice coefficients,<br />

7.3 SOLVING RECURRENCES 325<br />

+ . . . + al P? * (7.27)<br />

We will show that every rational function R(z) such that R(0) # 00 can be<br />

expressed in the <strong>for</strong>m<br />

R(z) = S(z) t T(z), (7.28)<br />

where S(z) has the <strong>for</strong>m (7.26) and T(z) is a polynomial. There<strong>for</strong>e there is a<br />

closed <strong>for</strong>m <strong>for</strong> the coefficients [z”] R(z). Finding S(z) and T(z) is equivalent<br />

to finding the “partial fraction expansion” of R(z).<br />

Notice that S(z) = 00 when z has the values l/p,, . . . , l/pi. There<strong>for</strong>e<br />

the numbers pk that we need to find, if we’re going to succeed in expressing<br />

R(z) in the desired <strong>for</strong>m S(z) + T(z), must be the reciprocals of the numbers<br />

&k where Q(ak) = 0. (Recall that R(z) = P(z)/Q(z), where P and Q are<br />

polynomials; we have R(z) = 00 only if Q(z) = 0.)<br />

Suppose Q(z) has the <strong>for</strong>m<br />

Q(z) = qo+q1z+~~~+q,z”‘, where qo # 0 and q,,, # 0.<br />

The “reflected” polynomial<br />

QR(z) = qoP+ q,z"-' +...f q,,,<br />

has an important relation to Q (2):<br />

QR(4 = qo(z - PI 1. . . (2 - P,)<br />

w Q(z) = qo(l -PIZ)...(~ -P~z)<br />

Thus, the roots of QR are the reciprocals of the roots of Q, and vice versa.<br />

We can there<strong>for</strong>e find the numbers pk we seek by factoring the reflected polynomial<br />

QR(z).<br />

For example, in the Fibonacci case we have<br />

Q(z) = 1 -2-z’; QR(z) = z2-z-l.<br />

The roots of QR ca.n be found by setting (a, b, c) = (1, -1, -1) in the quadratic<br />

<strong>for</strong>mula (-b II: da)/2a; we find that they are<br />

l+ds<br />

+=2<br />

1-d<br />

a n d $ = 2<br />

There<strong>for</strong>e QR(z) = (z-+)(2-$) and Q(z) = (1 -+z)(l -i$z).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!