Ratios: Analysts like to look at and
classify series in terms of the ratios of consecutive coefficients:
an + 1/an.
The reason for this, is that this is something like a discrete version
of
We think of the shift operator T(an) = an + 1 as the discrete version of the
The shift operator in quantum mechanics is the exponential of the momentum operator
f(x + h) − f(x) ≈ hf′(x)
f(x + h)/f(x) ≈ hf′(x)/f(x) + 1
A series ∑nanxn is called geometric if there is a constant R such that $$ \dfrac{a_{n+1}}{a_{n}} = R, \quad \forall n $$
A series ∑nanxn is called hypergeometric if $$ \dfrac{a_{n+1}}{a_{n}} = \dfrac{A(n)}{B(n)}, \quad \forall n $$
is a rational function in n.
Assuming we're working over an algebraically closed field, we factor the numerator and denominator as
A(n) = (n − a1)…(n − ap) and B(n) = (n − b1)…(n − bq)
Then since $a_{n+1} = \dfrac{A(n)}{B(n)} a_{n }= \dfrac{A(n)}{B(n)} \ldots \dfrac{A(0)}{B(0)} a_{0} $
We get terms like (n − a1)(n − 1 − a1)…(n − n − a1)
Convention: The standard way to write
these things is to factor it as (n + a1)… (so
you just take my a1 → −a1)
and then write
(a + n)(a + (n − 1))…a = (a)n This is called the Pochhammer symbol.
Then all together we can write our hypergeometric series as
$$ F_{p,q}(\bar{a} , \bar{b}, z) = \sum \dfrac{a_{0} (a_{1})_{k} \ldots (a_{p})_{k}}{b_{0} (b_{1})_{k} \ldots (b_{q})_{k} } z^{k} $$
Convention: Some other conventions are
to assume a0, b0 = 1
and that the first non-trivial root of B(n) is b1 = 1, so that (b1)k = (1)k = k!
, allowing you to have something like zk/k! ,
making your series look like an exponential type of thing. In this case,
they'll assume you have q + 1
roots on the bottom, so you write
$$ F_{p,q}(\bar{a} , \bar{b}, z) = \sum \dfrac{(a_{1})_{k} \ldots (a_{p})_{k}}{(b_{1})_{k} \ldots (b_{q})_{k} } \dfrac{z^{k}}{k!} $$
A series ∑nanxn is called hypergeometric if $$ \dfrac{a_{n+1}}{a_{n}} = A(n) $$
is a polynomial function in n.