← Back to Math Roadmap

Fractal Geometry

Self-similarity across scales, fractional dimensions, and deterministic chaos — the geometry of roughness.

What Is a Fractal?

▼
A fractal is a shape that exhibits self-similarity: zooming in reveals smaller copies of the whole. Unlike smooth curves from classical geometry, fractals are infinitely detailed and rough. Benoit Mandelbrot coined the term in 1975 and showed that fractals describe natural shapes — coastlines, clouds, mountains, trees, blood vessels — far better than Euclidean geometry ever could. The fractal dimension is a number (often non-integer) that quantifies how completely a fractal fills space. A smooth line has dimension one. The Koch snowflake (which adds smaller and smaller triangles infinitely) has dimension approximately 1.26. The Sierpinski triangle has dimension approximately 1.585. A space-filling curve like the Hilbert curve approaches dimension two.
Fractal (similarity) dimension. N is the number of self-similar pieces. r is the scaling factor (each piece is r times the size of the whole).

Classic Fractals and Their Dimensions

▼
The Koch snowflake starts with an equilateral triangle. In each iteration, replace every line segment's middle third with two segments forming a smaller equilateral triangle. After infinite iterations, the boundary has infinite length but finite area! Dimension D = log(4)/log(3) ≈ 1.26 — between a line (D=1) and a plane (D=2). The Sierpinski triangle starts with a filled triangle, then removes the central inverted triangle at each iteration. Dimension D = log(3)/log(2) ≈ 1.585. The Cantor set repeatedly removes the middle third of each interval. What remains is an uncountably infinite set of points with total length zero. Dimension D = log(2)/log(3) ≈ 0.63 — between a point (D=0) and a line (D=1). These fractional dimensions characterize how completely the fractal fills space.

The Mandelbrot set is defined by iterating z_{n+1} = z_n² + c for complex c. If z stays bounded, c is in the set (black). If z escapes to infinity, the escape speed determines the color. The boundary has fractal dimension exactly 2 — so convoluted it fills area. Zooming in anywhere reveals infinite self-similar complexity: spirals, seahorses, and miniature Mandelbrot copies. The related Julia sets fix c and vary the starting z, producing an infinite family of fractals, each corresponding to a point in the Mandelbrot set.
Mandelbrot iteration. For each complex number c, iterate from z₀=0. If |z_n| stays bounded (never exceeds 2), c belongs to the Mandelbrot set. The iteration count until escape determines the color.

The Mandelbrot Set

▼
The Mandelbrot set is the most famous fractal, often called the "thumbprint of God" for its infinite, psychedelic intricacy. It is defined by a deceptively simple iterative formula: start with z equals zero, then repeatedly compute z equals z squared plus c, where c is a complex number. If z stays bounded (does not escape to infinity) after many iterations, c belongs to the Mandelbrot set and is colored black. If z escapes, the number of iterations until escape determines the color. The boundary of the set is infinitely complex — it has a fractal dimension of two, meaning it is so convoluted that it essentially fills an area. Zooming in anywhere on the boundary reveals infinitely new patterns of spirals, seahorses, and miniature Mandelbrot copies.

Chaos Theory and the Butterfly Effect

▼
Fractals are deeply connected to chaos theory. A chaotic system is deterministic (no randomness) but exhibits extreme sensitivity to initial conditions — the butterfly effect: a butterfly flapping its wings in Brazil could theoretically cause a tornado in Texas. The Lyapunov exponent quantifies this sensitivity. A positive Lyapunov exponent indicates chaos — nearby trajectories diverge exponentially. The logistic map (x next equals r times x times one minus x) is the simplest system that exhibits chaos. As the parameter r increases, the system undergoes period-doubling bifurcations until reaching full chaos at r approximately 3.57. The Feigenbaum constant (approximately 4.669) describes the universal rate at which these bifurcations occur and appears in many different physical systems.

Fractals in Nature and Technology

▼
Fractals in Nature and Technology
Natural fractals: coastlines (Britain's coastline length depends on measurement scale), trees (branching patterns), lungs (maximizing surface area), lightning (dielectric breakdown), snowflakes, and galaxies. Applications: computer graphics (generating realistic landscapes), antenna design (fractal antennas are compact and multi-band), image compression (fractal compression), medicine (cancerous blood vessel networks have different fractal dimensions than healthy ones), and financial market analysis (fractal market hypothesis).

Fractal Generator

▼
Watch a fractal pattern emerge through iteration. Adjust the recursion depth to see more detail appear at finer scales.