generating function (analytic)
Suppose you have an infinite sequence of numbers a_0, a_1, a_2, ... — say the counts of something at each size. A generating function packages the whole sequence into a single object by making the numbers the coefficients of a power series. It is like hanging the sequence on a clothesline of powers of x, so that the entire sequence becomes one analytic function you can manipulate.
The ordinary generating function of (a_n) is the power series G(x) = sum_{n=0}^infinity a_n x^n. When this series has positive radius of convergence, G is an honest analytic function, and operations on the function mirror operations on the sequence: multiplying two generating functions corresponds to convolving (Cauchy-product) the sequences, differentiating shifts and scales indices, and so on. Closed-form expressions for G then encode facts about the a_n.
The analytic viewpoint adds real power beyond formal bookkeeping. Because G is analytic, its singularities — the nearest points where it blows up — control the growth rate of the coefficients a_n through the radius of convergence and finer asymptotic analysis. One honest caveat: a generating function is sometimes used purely formally, as a bookkeeping device with no convergence claimed; the analytic theory applies precisely when the series does converge near 0.
The Fibonacci numbers F_0=0, F_1=1, F_{n}=F_{n-1}+F_{n-2} have generating function sum F_n x^n = x/(1 - x - x^2), valid for |x| < 1/phi where phi is the golden ratio. The denominator's smaller root sets the radius and dictates the F_n grow like phi^n.
Fibonacci numbers encoded by a rational generating function.