A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic
Asymptotics
http://ac.cs.princeton.edu
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4a.CARM.Roadmap
Analytic combinatorics overview

specification
A. SYMBOLIC METHOD

1. OGFs

2. EGFs GF
equation

3. MGFs
SYMBOLIC METHOD
B. COMPLEX ASYMPTOTICS
asymptotic
4. Rational & Meromorphic estimate

5. Applications of R&M
COMPLEX ASYMPTOTICS
6. Singularity Analysis
desired
7. Applications of SA result !

8. Saddle point

3
Starting point

The symbolic method supplies generating functions that vary widely in nature.
√ −
+ − + + + ... +
( )= ( )= ( )= ( )=
− − − ... −

( )= ( )= ln ( )= + /
( )( )...( )

Next step: Derive asymptotic estimates of coefficients.

[ ] ( ) [ ] ( ) [ ] ( ) +
[ ] ( )=
(ln )
/ / /
[ ] ( ) [ ] ( ) = ln [ ]( )
!

Classical approach: Develop explicit expressions for coefficients, then approximate

Analytic combinatorics approach: Direct approximations.

4
Starting point
Catalan trees Derangements
Construction G = ○ × SEQ( G ) Construction D = SET (CYC>1( Z ))

ln
OGF equation ( )= EGF equation ( )=
( )
√
+ − =
Explicit form of OGF ( )= Explicit form of EGF

!" # ( )
Expansion ( )=− (− ) Expansion ( )=
!
≥

( )
Explicit form of coefficients = Explicit form of coefficients =
!

Approximation Approximation

/ /
Problem: Explicit forms can be unwieldy (or unavailable). ( + + / ! + ... + / !)

Opportunity: Relationship between asymptotic result and GF.

5
Analytic combinatorics overview
To analyze properties of a large combinatorial structure:
Ex. Derangements
Specification
1. Use the symbolic method (lectures 1 and 2).
D = SET (CYC>1( Z ))
• Define a class of combinatorial objects.
• Define a notion of size (and associated GF)
Symbolic transfer
• Use standard constructions to specify the structure.
• Use a symbolic transfer theorem.
Result: A direct derivation of a GF equation. −
GF equation ( )=
−

2. Use complex asymptotics (starting with this lecture).

Analytic transfer
• Start with GF equation.
• Use an analytic transfer theorem.
Result: Asymptotic estimates of the desired properties. ∼
Asymptotics

6
A shift in point of view

Specification

generating functions are treated as formal objects

Symbolic transfer
formal analytic
object! object!

GF equation
GF

Analytic transfer

generating functions are treated as analytic objects

Asymptotics

7
GFs as analytic objects (real)

Q. What happens when we assign real values to a GF?

coefficients are positive

so f(x) is positive (0, 1)

( )= (1, −1)
singularity continuation

A. We can use a series representation (in a certain interval) that allows us to extract coefficients.

= + + + + ... ≤ < / [ ]( )=
−
Useful concepts:
Differentiation: Compute derivative term-by-term where series is valid. ( )= + + + ...
Singularities: Points at which series ceases to be valid.
Continuation: Use functional representation even where series may diverge. ( )=−
8
GFs as analytic objects (complex)

Q. What happens when we assign complex values to a GF?

stay tuned for

interpretation
( )= of plot
singularity

A. We can use a series representation (in a certain domain) that allows us to extract coefficients.

Same useful concepts:

Differentiation: Compute derivative term-by-term where series is valid.
Singularities: Points at which series ceases to be valid.
Continuation: Use functional representation even where series may diverge.
9
GFs as analytic objects (complex)

Q. What happens when we assign complex values to a GF?

stay tuned for

interpretation
( )= of plot
singularity

A. A surprise!
Serendipity
is not
an accident
Singularities provide full information on growth of GF coefficients!

“Singularities provide a royal road to coefficient asymptotics.”

10
General form of coefficients of combinatorial GFs

First principle of coefficient asymptotics

subexponential The location of a function’s singularities dictates
factor the exponential growth of its coefficients.
[ ] ( )= θ( )
exponential Second principle of coefficient asymptotics
growth factor
The nature of a function’s singularities dictates
the subexponential factor of the growth.

Examples (preview): GF singularities exponential subexp.

GF
type location nature growth factor

strings with −
no 00 ( )= rational /φ, /φ̂ pole φ √
− −

( )= −
derangements meromorphic 1 pole 1N
√
+ − square
Catalan trees ( )= analytic / 4N
root
11
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4a.CARM.Roadmap
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4b.CARM.Complex
Theory of complex functions

Quintessential example of the power of abstraction.

are complex
numbers
Start by defining i to be the square root of −1 so that i 2 = −1 real ?

Continue by exploring natural definitions of basic operations

• Addition
• Multiplication
• Division
1+i
• Exponentiation
• Functions
• Differentiation
• Integration

14
Standard conventions

Correspondence with points in the plane

= +
(x, y) represents
x z = x + iy
real part

|z| y
imaginary part

absolute value | | +

conjugate ¯=

(x, −y) represents

Quick exercise: ¯=| | z = x − iy

15
Basic operations

Natural approach: Use algebra, but convert i 2 to −1 whenever it occurs

Addition
( + )+( + )=( + )+( + )

Multiplication
( + ) ( + )= + + +
=( )+( + )

Division
¯
= =
+ + | |

Exponentiation?

16
Analytic functions

Definition. A function f (z ) defined in Ω is analytic at a point z0 in Ω iff for z in an open disc in

Ω centered at z0 it is representable by a power-series expansion ( )= ( )

Examples:

= + + + + + ... is analytic for |z| < 1 .

+ + + + + ... is analytic for |z| < ∞ .

! ! ! !

17
Complex differentiation

Definition. A function f (z ) defined in a region Ω is holomorphic or complex-differentiable at a

( + ) ( )
point z0 in Ω iff the limit ( ) = lim exists, for complex δ.

Note: Notationally the same as for reals, but much stronger—the value is independent of the way that δ approaches 0.

Theorem. Basic Equivalence Theorem. For purposes of

this lecture:
A function is analytic in a region Ω iff it is complex-differentiable in Ω. Axiom 1.

Useful facts:
• If function is analytic (complex-differentiable) in Ω, it admits derivatives of any order in Ω.
• We can differentiate a function via term-by-term differentiation of its series representation.
• Taylor series expansions ala reals are effective.

18
Taylor's theorem
immediately gives power series expansions for analytic functions.

+ + + + + ...
! ! ! !

sin + + ...
! ! ! !

cos + + ...
! ! !

= + + + + + ...

19
Euler's formula

Evaluate the exponential function at iθ

( ) ( ) ( ) ( )
= + + + + + ... + + ...
! ! ! ! !

= + + + + + + ...
! ! ! ! ! ! ! ! !
i 2 = −1 i 3 = −i i4 = 1

“Our jewel . . . one of the most remarkable, almost

= cos + sin
astounding, formulas in all of mathematics”
Euler's formula — Richard Feynman, 1977

20
Polar coordinates

Euler's formula gives another correspondence between complex numbers and points in the plane.

r cosθ (x, y)
= cos + sin
r r sinθ
θ

Conversion functions defined for any complex

number x + iy :

• absolute value (modulus) = +

• angle (argument) = arctan

21
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4b.CARM.Complex
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4c.CARM.Rational
Rational functions
are complex functions that are the ratio of two polynomials.

√ −
+ − + + + ... +
( )= ( )= ( )= ( )=
− − − ... −

( )= ( )= ln ( )= + /
( )( )...( )

Approach:

• Use partial fractions to expand into terms for which coefficient extraction is easy.

• Focus on the largest term to approximate.

[Same approach as for reals, but takes complex roots into account.]

24
Extracting coefficients from rational GFs

Factor the denominator and use partial fractions to expand into sum of simple terms.

Example 1.
Rational GF ( )= ( )
(distinct roots) +

Factor denominator =
( )( )
Use partial fractions: ( )= +
Expansion must be of the form

Cross multiply + =
and solve for coefficients.
+ =

Solution is c0 = 1 and c1=−1 ( )=

Extract coefficients. =[ ] ( )=
25
Extracting coefficients from rational GFs

Factor the denominator and use partial fractions to expand into sum of simple terms.

Example 2.
Rational GF ( )=
(multiple roots) +

Factor denominator =
( + )( )

Use partial fractions:

Expansion must be of the form
( )= + +
+ ( )
+ + =
Cross multiply + =
and solve for coefficients.
=

Solution is
c0 = −2/9, c1 = −1/9, and c2= 3/9
( )= +
+ ( )
Extract coefficients. =[ ] ( )= ( ( ) + + )
26
Approximating coefficients from rational GFs

When roots are real, only one term matters.

( )= +
+ ( )

= ( ( ) + + )

smaller roots give

( + ) exponentially smaller terms

multiplicity 3 gives terms

of the form n2βn, etc.

27
Extracting coefficients from rational GFs

Factor the denominator and use partial fractions to expand into sum of simple terms.

Example 3.
Rational GF ( )=
(complex roots) +

Factor denominator = =
( )( + ) ( + )
Use partial fractions: ( )= +
Expansion must be of the form +

Cross multiply + =
and solve for coefficients. =

Solution is ( )= +
c0 = c1 = 1/2 +

Extract coefficients. [ ] ( )= ( +( ) )= ( +( ) )

1, 0, -1, 0, 1, 0, -1, 0, 1... 28

Extracting coefficients form rational GFs (summary)

Theorem. Suppose that g(z) is a polynomial of degree t with roots β1, β2,..., βr and let mi
denote the multiplicity of βi for i from 1 to r. If f (z) is another polynomial with no roots in
common with g(z), and g(0)≠0 then

( ) ! ! !
[ ] = β + β + ... + β
( )
≤< ≤< ≤<

Notes:
+ + ... + =

• There are t terms, because m1 + m2 + ... + mr = t.

• The t constants cij depend upon f.

• Complex roots introduce periodic behavior.

29
AC transfer theorem for rational GFs (leading term)

Theorem. Assume that a rational GF f (z)/g(z) with f (z) and g(z) relatively prime and g(0)≠0
has a unique pole of smallest modulus 1/β and that the multiplicity of β is ν. Then
( ) ( ) ( / )
[ ] = ( )( / )
( )
typical case

A(z) = f (z)/g(z) 1/β ν C [zN ]A(z)

Examples.
( ) ( / )
1/2 2 =
+

( / )
φ 1 =
( / )

+ + +
1.9276... 1 1.09166...

30
Computer algebra solution

Transfer theorem amounts to an algorithm that is embodied in many computer algebra systems.

31
Classic example: Algorithm for solving linear recurrences

pp. 157–158

32
AC example with rational GFs: Patterns in strings

B4, the class of all binary strings with no 04 see Lecture 1

Specification
B4 = Z<4 (E + Z1B4)

Symbolic transfer

( )=( + + + )( + ( ))
GF equation + + +
=
− − − −

Analytic transfer

. .
Asymptotics ∼ β = . β= .

Many more examples to follow (next lecture)

33
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4c.CARM.Rational
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4d.CARM.Analytic
Analytic functions

Definition. A function f (z ) defined in Ω is analytic at a point z0 in Ω iff for z in an open disc in

Ω centered at z0 it is representable by a power-series expansion ( )= ( )

Definition. A singularity is a point where a function ceases to be analytic.

Example: = analytic at 0

= =
( )
+
= ( ) analytic everywhere but z = 1

36
Analytic functions

function region of meromorphicity

+ + everywhere

everywhere but z = 0

Definition. A function f (z ) defined in

Ω is analytic at a point z0 in Ω iff for ( )= everywhere but z = 1

z in an open disc in Ω centered at z0

everywhere but z = ±i
it is representable by a power-series +
expansion ( )= ( )
+ / everywhere
( )=

( )= everywhere but z = 1, 1/2, 1/3, ...

( )( )...( )
√
+ − everywhere but z = 1/4
( )=

( )= everywhere but z = ln2 ± 2πki

( )= ln everywhere but z = 1
37
Aside: computing with complex functions
is an easy exercise in object-oriented programming.

public class Complex

{ public interface ComplexFunction
private final double re; // real part {
private final double im; // imaginary part public Complex eval(Complex z);
}
public Complex(double real, double imag)
{
re = real;
im = imag; public class Example implements ComplexFunction
} {
public Complex eval(Complex z)
public Complex plus(Complex b) { // {1 \over 1+z^3}
{ Complex one = new Complex(1, 0);
Complex a = this;
Complex d = one.plus(z.times(z.times(z)));
double real = a.re + b.re;
return d.reciprocal();
double imag = a.im + b.im;
}
return new Complex(real, imag);
}
}

public Complex times(Complex b)

{ Design choice: complex numbers are immutable
Complex a = this;
double real = a.re * b.re - a.im * b.im;
• create a new object for every computed value
double imag = a.re * b.im + a.im * b.re;
return new Complex(real, imag);
} • object value never changes

... [Same approach as for Java strings.]

}
38
Aside (continued): plotting complex functions
is also an easy (and instructive!) programming exercise.
public class Plot2Dez public class Example implements ComplexFunction
{ {
public static void show(ComplexFunction f, int sz) public Complex eval(Complex z)
{ { // {1 \over 1+z^3}
StdDraw.setCanvasSize(sz, sz); Complex one = new Complex(1, 0);
StdDraw.setXscale(0, sz); Complex d = one.plus(z.times(z.times(z)));
StdDraw.setYscale(0, sz); return d.reciprocal();
double scale = 2.5; }
for (int i = 0; i < sz; i++) public static void main(String[] args)
for (int j = 0; j < sz; j++) { Plot2D.show(new Example(), 512); }
{ }
double x = ((1.0*i)/sz - .5)*scale;
double y = ((1.0*j)/sz - .5)*scale; our convention:
Complex z = new Complex(x, y); plots are in the 2.5 by 2.5 square
double val = f.eval(z).abs()*10; centered at the origin
int t;
if (val < 0) t = 255;
else if (val > 255) t = 0; arbitrary factor
else t = 255 - (int) val; to emphasize growth
Color c = new Color(t, t, t); +
StdDraw.setPenColor(c);
StdDraw.pixel(i, j);
}
Color c = new Color(0, 0, 0);
StdDraw.setPenColor(c); singularities
StdDraw.line(sz/2, 0, sz/2, sz); (where |f | → ∞)
StdDraw.line(0, sz/2, sz, sz/2);
darkness of pixel at (x, y)
StdDraw.show();
is proportional to |f (x + iy )|
}
} 39
Entire functions (analytic everywhere)

+ +

+ /

our convention:
highlight the 2.5 by 2.5 square
centered at the origin
when plotting a bigger square

40
Plots of various rational functions

+ + +
− − − −

( )( )( )( )

41
Complex integration

(2, 3)
! ! −
Starting point: = ( + ) = + =
Change variables to
L !−
convert to real !
= − ! = +
integrals.
(2, −1)

Augustin-Louis Cauchy
1789-1857

Amazing facts:

• The integral of an analytic function around a loop is 0.

• The coefficients of an analytic function can be extracted via complex integration

Analytic combinatorics context: Immediately gives exponential growth for meromorphic GFs
42
Integration examples
! ! "
"
= + = " + =− +
Ex 1. Integrate f (z) = z on a rectangle − −
! ! − "−
(−4, 3)
L1 (2, 3) "
= ( + ) = − " = + = + =
R
! ! − "−
"
L4 L2 = − = " − = −
! ! "
"
(−4, −1) (2, −1)
= (− + ) =− − " =− −
− −
L3
! !
= =− + + + + − − − = (!)
+ + +

= =

! ! θ!
C θ !
Ex 2. Integrate f (z) = z on a circle centered at 0 = θ= ! = ( − )=

! !
Ex 3. Integrate f (z) = 1/z on a circle centered at 0 = θ=

43
Integration examples

Ex 4. Integrate f (z) = zM on a circle centered at 0

! !
+ ( + )θ
C = θ = =

!
!
θ=
=−
= !
"= − ( + )θ "
( + )θ "
θ= " = ( − )=
( + ) ( + )

Ex 5. Integrate f (z) = (z−s)M on a circle centered at s

! !
Cs ( − ) + ( + )θ = =
= θ

s !
=−
=
"= −

44
Null integral property

Theorem. (Null integral property). For purposes of

this lecture:
If f (z) is analytic in Ω then ( ) = for any closed loop λ in Ω. Axiom 2.

! !
Ex. f (z) = z = =

Homotopic: Paths that can

be continuously deformed
Equivalent fact: ( ) = ( ) for any homotopic paths α and β in Ω. into one another.

α α
β β

( ) = ( ) = ( )
+
45
Deep theorems of complex analysis

Complex
Differentiability

Analyticity

Null Integral
Property

Appendix C
pp. 741-743

46
Cauchy’s coefficient formula

Theorem. If f (z) is analytic and λ is a closed +loop in a region Ω that contains 0, then

[ ] ( ) = ( ) +

Proof.

• Expand f : ( )= + + + + ...
0
• Deform λ to a circle centered at 0

• Integrate: ( ) +
= +
+ ... + + + + + + ...

= See integration example 4

AC context: provides transfer theorems for broader class of complex functions: meromorphic functions (next).

47
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4d.CARM.Analytic
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4e.CARM.Meromorphic
Meromorphic functions
are complex functions that can be expressed as the ratio of two analytic functions.

Note: All rational functions are meromorphic.

√ −
+ − + + + ... +
( )= ( )= ( )= ( )=
− − − ... −

( )= ( )= ln ( )= + /
( )( )...( )

Approach:

• Use contour integration to expand into terms for which coefficient extraction is easy.

• Focus on the largest term to approximate.

[Same approach as for rationals, resulting in a more general transfer theorem.]

50
Meromorphic functions

Definition. A function h(z ) defined in Ω is meromorphic at z0 in Ω iff for z in a neighborhood

of z0 with z ≠ z0 it can be represented as f (z)/g(z), where f (z) and g(z) are analytic at z0.

Useful facts:

• A function h(z) that is meromorphic at z0 admits an expansion of the form

( )= + ... + + + + ( )+ ( ) + ...
( ) ( ) ( )
and is said to have a pole of order M at z0. Proof sketch: If z0 is a zero of g(z) then g(z) = (z − z0)M G(z).
Expand the analytic function f (z)/G(z) at z0.

• The coefficient h−1 is called the residue of h(z) at z0, written ( ).

=
• If h(z) has a pole of order M at z0, the function (z − z0)M h(z) is analytic at z0.

A function is meromorphic in Ω iff it is analytic in Ω except for a set of isolated singularities, its poles.
51
Meromorphic functions

function region of meromorphicity

+ + everywhere

everywhere but z = 0
Definition. A function h(z ) defined in
Ω is meromorphic at z0 in Ω iff for z
in a neighborhood of z0 with z ≠ z0 it ( )= everywhere but z = 1
can be represented as f (z)/g(z),
where f (z) and g(z) are analytic at z0.
everywhere but z = ±i
+

( )= everywhere but z = 1, 1/2, 1/3, ...

( )( )...( )

( )= everywhere but z = ln2 ± 2πki

52
Plots of various meromorphic functions

+ + +
− − − −

( )( )( )( )

53
Integrating around a pole

Lemma. If h(z) is meromorphic and λ is a closed +loop with s

a single pole s of h inside, then ( ) = ( )

=

!
C
Ex. f (z) = 1/z, pole at 0 with residue 1. =

Proof.

• Expand h: ( )= + ... + + + ( )+ ( ) + ...

( ) ( )
s
• Deform λ to a circle centered at s that contains no other poles

• Integrate: ( ) = + ... + + + ( )+ ( ) + ...

( ) ( )

= See integration example 5

Significance: Connects local properties of a function (residue at a point)

to global properties elsewhere (integral along a distant curve).
54
Residue theorem

Theorem. If h(z) is meromorphic and λ is a closed +loop in Ω, then

( ) = ( )
=
λ
where S is the set of poles of h(z) inside Ω

Proof (sketch).

• Consider small circles Cs centered at each pole.

λ*
• Define a path λ* that follows λ but travels in, around, and out each Cs.

• Poles are all outside λ* so integral around λ* is 0.

• Paths in and out cancel, so ( ) = ( ) ( ) =

• By the single-pole lemma ( ) = ( )

=

55
Extracting coefficients from meromorphic GFs

Theorem. Suppose that h (z ) is meromorphic in the closed disc |z | ≤ R; analytic

at z = 0 and all points |z | = R; and that α1, ... αm are the poles of h (z ) in R. Then
( ) ( ) ( )
[ ] ( )= + + ... + +

where p1, ..., pm are polynomials with degree α1−1, ..., αm −1, respectively.
R
α3
α1
α2
Proof sketch: αm

• Consider the integral = ( ) +

| |=
Ex. If αi is order 1

( ) ( ) ( ) ( )
= ( )
• By the residue theorem = + = + ... +
( )
+
= +
= = ( )
=
< | ( )| < | |= +
• By direct bound
Constant. May depend on R, but not N. 56
Complex roots

Q. Do complex roots introduce complications in deriving asymptotic estimates of coefficients?

A. YES: all poles closest to the origin contribute to the leading term.

Prime example: Nth roots of unity = exp( ) = cos( ) + sin( ) <

all are distance 1 from origin with ( ) =

Rational GF example earlier in this lecture.

[ ] = , , , , , , ,...
+
57
Complex roots

Q. Do complex roots introduce complications in deriving asymptotic estimates of coefficients?

A. NO, for combinatorial GFs, if only one root is closest to the origin.

+ + +
Pringsheim’s Theorem. If h (z) can be represented as a series
− − − −
expansion in powers of z with non-negative coefficients and radius
of convergence R, then the point z = R is a singularity of h (z).

smallest positive real root

Implication: Only the smallest positive real root matters if no others have the same magnitude.

If some do have the same magnitude,

complicated periodicities can be present.
See "Daffodil Lemma" on page 266.
58
AC transfer theorem for meromorphic GFs (leading term)

Theorem. Suppose that h (z )= f (z)/g(z) is meromorphic in |z | ≤ R and analytic both at z = 0

and at all points |z | = R. If α is a unique closest pole to the origin of h (z ) in R, then α is real
( ) ( )
and [ ] where M is the order of α, =( ) ( )( )
and β = 1/α.
( )

Proof sketch for M = 1:

elementary from Pringsheim’s and
• Series expansion (valid near α): ( )= + + ( )+ ( ) + ...
coefficient extraction theorems
• One way to calculate constant: = lim ( ) ( )

• Approximation at α: ( ) = =
/
See next slide for calculation of c and M > 1.

Notes:
• Error is exponentially small (and next term may involve periodicities due to complex roots).
• Result is the same as for rational functions.

59
Computing coefficients for a meromorphic function h(z) = f(z)/g(z) at a pole α

If α is of order 1 then [ ] ( ) +
= lim ( ) ( )

( )( ) ( ) ( ) ( ) ( )
To calculate h−1: lim ( ) ( ) = lim = lim =
( ) ( ) ( )

If α is of order 2 then [ ] ( ) +
= lim ( ) ( )

Series expansion (valid near α): ( )= + + + ( )+ ( ) + ...

( )
( + )
Approximation at α: ( ) = =
( ) ( / )

( ) ( ) ( ) ( ) ( )( )
To calculate h−2: lim ( ) ( ) = lim = lim
( ) ( )
( ) ( ) ( ) ( )+ ( ) ( )
= lim =
( ) ( )

( )
If α is of order M then [ ] ( ) ( ) ( )( )
60
Bottom line

Specification

Symbolic transfer

GF equation Analytic transfer for meromorphic GFs: f (z)/g (z) ~ c βN

• Compute the dominant pole α (smallest real with g(z) = 0).
• (Check that no others have the same magnitude.)
Analytic transfer
• Compute the residue h−1 = −f (α)/g' (α). Not order 1 if g'(α) = 0.
• Constant c is h−1 /α. Adjust to (slightly) more
complicated order M case.
• Exponential growth factor β is 1/α
Asymptotics

61
AC transfer for meromorphic GFs

Analytic transfer for meromorphic GFs: f (z)/g (z) ~ c βN

• Compute the dominant pole α (smallest real with g(z) = 0).
• (Check that no others have the same magnitude.)
• Compute the residue h−1 = −f (α)/g' (α).
• Constant c is h−1 /α.
• Exponential growth factor β is 1/α

h(z) = f (z)/g(z) α h−1 [zN ]h(z)

Examples. ˆ=
ˆ ˆ
ˆ= =
( + ˆ) +
=

/ /
1

62
AC example with meromorphic GFs: Generalized derangements
DM, the class of all permutations
see Lecture 2
Specification
with no cycles of length ≤ M
DM = SET(CYC>M(Z )

Symbolic transfer

− − − −...
GF equation ( )=
−

Analytic transfer

Asymptotics [ ] ( )

Many, many more examples to follow (next lecture)

63
General form of coefficients of combinatorial GFs (revisited)

First principle of coefficient asymptotics

subexponential The location of a function’s singularities dictates
factor the exponential growth of its coefficients.
[ ] ( )= θ( )
exponential Second principle of coefficient asymptotics
growth factor
The nature of a function’s singularities dictates
the subexponential factor of the growth.

When F(z) is a meromorphic function f (z)/g (z)

• If the smallest real root of g (z) is α then the exponential growth factor is 1/α.
• If α is a pole of order M, then the subexponential factor is cNM−1.

64
Parting thoughts

“Despite all appearances, generating functions belong to algebra, not analysis”

— John Riordan, 1958

“Combinatorialists use recurrences, generating functions, and such transformations as the Vandermonde convolution;
Others, to my horror, use contour integrals, differential equations, and other resources of mathematical analysis”

— John Riordan, 1968

???
?
( ) ( )
[ ] =[ ] =
! !

/ /
( ) ( ) ( )
[ ] =[ ] = ...
! ! !
65
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

II.4e.CARM.Meromorphic
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic functions
Analytic
Combinatorics • Roadmap
Philippe Flajolet and
• Complex functions
• Rational functions
Robert Sedgewick
OF

• Analytic functions and complex integration

• Meromorphic functions
C A MBRI D GE

http://ac.cs.princeton.edu

• Exercises II.4f.CARM.Exercises
Note IV.28

Supernecklaces

Warmup: A "supernecklace" of the 3rd type is a labelled cycle of cycles.

Draw all the supernecklaces of the 3rd type of size N for N = 1, 2, 3, and 4.

68
Assignments

1. Read pages 223-288 (Complex Analysis, Rational, and Meromorphic Functions) in text.
Usual caveat: Try to get a feeling for what's there, not understand every detail.

2. Write up solution to Note IV.28.

3. Programming exercises.

Program IV.1. Compute the percentage of permutations

having no singelton or doubleton cycles and compare with
the AC asymptotic estimate, for N = 10 and N = 20. .

Program IV.2. Plot the derivative of the supernecklace GF (see Note IV.28)
in the style of the plots in this lecture (see booksite for Java code).

69
 A N A LY T I C C O M B I N AT O R I C S

PART TWO

4. Complex Analysis,
Rational and Meromorphic
Asymptotics
http://ac.cs.princeton.edu