Complex quadratic polynomial

A complex quadratic polynomial is a quadratic polynomial whose coefficients are complex numbers.

Forms

When the quadratic polynomial has only one variable ( univariate ), one can distinguish its 4 main forms:

• The general form : $f(x) = a_2 x^2 + a_1 x + a_0 \qquad \,$ where $\qquad a_2 \ne 0$.
• The factored form used for logistic map $f_r(x) = r x ( 1-x ) \,$
• $f_{\theta}(x) = x^2 + e^{2 \pi \theta i} x \,$ which has an indifferent fixed point with multiplier $\lambda = e^{2 \pi \theta i} \,$ at the origin
• The monic and centered form, $f_c(x) = x^2 +c\,$

The monic and centered form has the following properties:

• It is the simplest form of a nonlinear function with one coefficient (parameter),
• It is an unicritical polynomial, i.e. it has one critical point,
• It is a centered polynomial (the sum of its critical points is zero),
• It can be postcritically finite, i.e. If the orbit of the critical point is finite. It is when critical point is periodic or preperiodic.
• It is a unimodal function,
• It is a rational function,
• It is an entire function.

Conjugation

Between forms

Since $f_c(x) \,$ is affine conjugate to the general form of the quadratic polynomial it is often used to study complex dynamics and to create images of Mandelbrot, Julia and Fatou sets.

When one wants change from $\theta\,$ to $c \,$ :

$c = c(\theta) = \frac {e^{2 \pi \theta i}}{2} \left(1 - \frac {e^{2 \pi \theta i}}{2}\right)$

When one wants change from $r\,$ to $c \,$ :

$c = c(r)\,=\,\frac{1- (r-1)^2}{4}$

With doubling map

There is semi-conjugacy between the dyadic transformation (here named doubling map) and the quadratic polynomial.

Family

The family of quadratic polynomials $f_c : z \to z^2 +c\,$ parametrised by $c \in \mathbb{C} \,$ is called :

• the Douady-Hubbard family of quadratic polynomials
• quadratic family

Map

The monic and centered form is typically used with variable $z\,$ and parameter $c\,$

$f_c(z) = z^2 +c.\,$

When it is used as an evolution function of the discrete nonlinear dynamical system:

$z_{n+1} = f_c(z_n) \,$

it is named quadratic map

$f_c : z \to z^2 + c. \,$

Notation

Here $f^n \,$ denotes the n-th iteration of the function $f \,$ not exponentiation

$f_c^n(z) = f_c^1(f_c^{n-1}(z)) \,$

so

$z_n = f_c^n(z_0). \,$

Because of the possible confusion it is customary to write $f^{\circ n}\,$ for the nth iterate of the function $f.\,$

Critical items

Critical point

A critical point of $f_c\,$ is a point $z_{cr} \,$ in the dynamical plane such that the derivative vanishes :

$f_c'(z_{cr}) = 0. \,$

Since

$f_c'(z) = \frac{d}{dz}f_c(z) = 2z$

implies

$z_{cr} = 0\,$

we see that the only (finite) critical point of $f_c \,$ is the point $z_{cr} = 0\,$.

$z_0$ is an initial point for Mandelbrot set iteration.

Critical value

A critical value $z_{cv} \$ of $f_c\,$ is the image of a critical point:

$z_{cv} = f_c(z_{cr}) \,$

Since

$z_{cr} = 0\,$

we have

$z_{cv} = c. \,$

So the parameter $c \,$ is the critical value of $f_c(z). \,$

Critical orbit

Dynamical plane with critical orbit falling into 3-period cycle

Dynamical plane with Julia set and critical orbit.

Dynamical plane : changes of critical orbit along internal ray of main cardioid for angle 1/6

Critical orbit tending to weakly attracting fixed point with abs(multiplier)=0.99993612384259

Forward orbit of a critical point is called a critical orbit. Critical orbits are very important because every attracting periodic orbit attracts a critical point, so studying the critical orbits helps us understand the dynamics in the Fatou set.

$z_0 = z_{cr} = 0\,$

$z_1 = f_c(z_0) = c\,$

$z_2 = f_c(z_1) = c^2 +c\,$

$z_3 = f_c(z_2) = (c^2 + c)^2 + c\,$

$... \,$

This orbit falls into an attracting periodic cycle

Critical sector

The critical sector is a sector of the dynamical plane containing the critical point .

Critical polynomial

$P_n(c) = f_c^n(z_{cr}) = f_c^n(0) \,$

so

$P_0(c)= 0 \,$

$P_1(c) = c \,$

$P_2(c) = c^2 + c \,$

$P_3(c) = (c^2 + c)^2 + c \,$

These polynomials are used for :

• finding centers of these Mandelbrot set components of period n. Centers are roots of n-th critical polynomials

$centers = \{ c : P_n(c) = 0 \}\,$

• finding roots of Mandelbrot set components of period n ( local minimum of $P_n(c) \,$ )
• Misiurewicz points

$M_{n,k} = \{ c : P_k(c) = P_{k+n}(c) \}\,$

Critical curves

Diagrams of critical polynomials are called critical curves.

These curves create skeleton of bifurcation diagram ( the dark lines )

Planes

w-plane and c-plane

One can use the Julia-Mandelbrot 4-dimensional space for a global analysis of this dynamical system.

In this space there are 2 basic types of 2-D planes :

• the dynamical ( dynamic ) plane, $f_c\,$-plane or c-plane
• the parameter plane or z-plane.

There is also another plane used to analyze such dynamical systems w-plane :

• the conjugation plane
• model plane

Parameter plane

Gamma parameter plane for complex logistic map $z_{n+1} = \gamma z_n \left(1 - z_n\right),$

The phase space of a quadratic map is called its parameter plane. Here:

$z0 = z_{cr} \,$ is constant and $c\,$ is variable.

There is no dynamics here. It is only a set of parameter values. There are no orbits on the parameter plane.

The parameter plane consists of :

• The Mandelbrot set
  • The bifurcation locus = boundary of Mandelbrot set
  • Bounded hyperbolic components of the Mandelbrot set = interior of Mandelbrot set = Connectedness locus
• The complement of the Mandelbrot set $M^c = \mathbb{\hat{C}}\setminus M$ = unbounded hyperbolic component

There are many different subtypes of the parameter plane

Dynamical plane

On the dynamical plane one can find:

• The Julia set
• The Filled Julia set
• The Fatou set
• Orbits .

The dynamical plane consists of:

• Fatou set
• Julia set

Here, $c\,$ is a constant and $z\,$ is a variable.

The two-dimensional dynamical plane can be treated as a Poincare cross-section of three-dimensional space of continuous dynamical system.

Derivatives

Derivative with respect to c

On parameter plane :

• $c$ is a variable
• $z_0 = 0$ is constant

The first derivative of $f_c^n(z_0)$ with respect to c is

$z_n' = \frac{d}{dc} f_c^n(z_0).$

This derivative can be found by iteration starting with

$z_0' = \frac{d}{dc} f_c^0(z_0) = 1$

and then replacing at every consecutive step

$z_{n+1}' = \frac{d}{dc} f_c^{n+1}(z_0) = 2\cdot{}f_c^n(z)\cdot\frac{d}{dc} f_c^n(z_0) + 1 = 2 \cdot z_n \cdot z_n' +1.$

This can easily be verified by using the chain rule for the derivative.

This derivative is used in the distance estimation method for drawing a Mandelbrot set.

Derivative with respect to z

On dynamical plane :

• $z$ is a variable
• $c$ is a constant

at a fixed point $z_0\,$

$f_c'(z_0) = \frac{d}{dz}f_c(z_0) = 2z_0$

at a periodic point z₀ of period p

$(f_c^p)'(z_0) = \frac{d}{dz}f_c^p(z_0) = \prod_{i=0}^{p-1} f_c'(z_i) = 2^p \prod_{i=0}^{p-1} z_i.$

It is used to check the stability of periodic (also fixed) points.

at nonperiodic point :

$z'_n\,$

This derivative can be found by iteration starting with

$z'_0 = 1 \,$

and then :

$z'_n= 2*z_{n-1}*z'_{n-1}\,$

This dervative is used for computing external distance to Julia set.

Schwarzian derivative

The Schwarzian derivative (SD for short) of f is :

$(Sf)(z) = \frac{f'''(z)}{f'(z)} - \frac{3}{2} \left ( \frac{f''(z)}{f'(z)}\right ) ^2$

See also

• Misiurewicz point
• Periodic points of complex quadratic mappings
• Mandelbrot set
• Julia set
• Milnor–Thurston kneading theory

External links

• M. Nevins and D. Rogers, "Quadratic maps as dynamical systems on the p-adic numbers"
• Wolf Jung : Homeomorphisms on Edges of the Mandelbrot Set. Ph.D. thesis of 2002