Fractals & Iterative Transformations

A fractal is a geometric object that displays self-similarity at every scale — zoom in on any piece and you see a miniature copy of the whole. Far from being mere curiosities, fractals arise naturally from a simple mathematical idea: apply the same transformation over and over again, infinitely many times. This process is called iteration, and when the transformations involved are linear (or affine) maps expressed as matrices, the entire theory falls squarely inside linear algebra. The key tool is the Iterated Function System (IFS). An IFS consists of a finite collection of contractive affine maps, each written in matrix form as f_i(x) = A_i x + b_i, where A_i is a 2×2 matrix and b_i is a translation vector. Starting from almost any initial set, repeatedly applying all the maps in random or systematic order causes the set of points to converge to a unique, highly structured limit called the attractor — and that attractor is the fractal. One of the most powerful connections between fractals and linear algebra is the geometric meaning of the determinant. For a 2×2 matrix A, |det(A)| tells you exactly how much the transformation scales areas. Because each IFS map is a contraction, |det(A_i)| < 1, and this fact controls how the fractal's area shrinks to zero and how its dimension is computed. Understanding both iteration and the determinant together gives you a deep, unified picture of why fractals look the way they do.

Key takeaways
  • Affine Maps in Matrix Form
  • Contraction and Convergence
  • The Determinant as an Area Scaling Factor
  • Fractal Dimension from the IFS
  • The Chaos Game Algorithm

Affine Maps in Matrix Form

An affine map in 2D is written f(x) = Ax + b, where A is a 2×2 matrix, x is a column vector [x, y]^T, and b is a translation vector. The matrix A handles rotation, scaling, and shearing; b shifts the result. For the Sierpiński triangle, three such maps are used, each with A_i = (1/2)I (a scaling by one-half) combined with different translations b_i that place the contracted copy in each corner of the triangle.

Contraction and Convergence

A map f is a contraction if there exists a constant 0 ≤ s < 1 such that ||f(x) − f(y)|| ≤ s||x − y|| for all points x, y. The number s is called the contraction ratio. Banach's Fixed-Point Theorem guarantees that any collection of contractions on a complete metric space has a unique fixed set — the attractor. This is why IFS fractals are well-defined mathematical objects that do not depend on your starting shape.

The Determinant as an Area Scaling Factor

When a linear transformation A is applied to a region, every area inside that region is multiplied by |det(A)|. If A = [[a, b],[c, d]], then det(A) = ad − bc. For a contraction, |det(A)| < 1, meaning area shrinks. For the Sierpiński triangle maps A_i = (1/2)I, det(A_i) = (1/2)^2 = 1/4. Each application reduces area by 75%. After n iterations, the total area of all 3^n triangles is (3 · (1/4))^n = (3/4)^n → 0 as n → ∞. The fractal has zero area, yet is not empty — a hallmark of fractional dimension.

Fractal Dimension from the IFS

The Hausdorff (fractal) dimension d of a self-similar fractal generated by N maps each with the same contraction ratio s satisfies N · s^d = 1, giving d = log(N) / log(1/s). For the Sierpiński triangle: N = 3 maps, s = 1/2, so d = log(3)/log(2) ≈ 1.585. This is between 1 (a curve) and 2 (a filled region), reflecting the fractal's intricate but incomplete area. The determinant condition |det(A_i)| = s^2 directly links the area-scaling viewpoint to the dimension formula.

The Chaos Game Algorithm

A practical algorithm to draw any IFS fractal: (1) Start at an arbitrary point x_0. (2) At each step, randomly choose one of the N maps f_i with probability p_i proportional to |det(A_i)|. (3) Set x_{n+1} = f_i(x_n). (4) Plot the point. After a few thousand iterations, the plotted points trace out the fractal attractor with remarkable accuracy. This works because the random orbit visits every region of the attractor proportionally to its 'weight', determined again by the determinants.

The Key Equation

xn+1=Aixn+bi,∣det⁡(Ai)∣=si 2<1\mathbf{x}_{n+1} = A_i \mathbf{x}_n + \mathbf{b}_i, \quad |\det(A_i)| = s_i^{\,2} < 1
  • A_i — Transformation matrix (dimensionless): A 2×2 matrix encoding rotation, scaling, and shearing for the i-th IFS map.
  • \mathbf{b}_i — Translation vector (dimensionless): A 2D column vector that shifts the contracted image to its correct position.
  • \det(A_i) — Determinant (dimensionless): Scalar equal to ad − bc for A = [[a,b],[c,d]]; its absolute value gives the area scaling factor of the map.
  • s_i — Contraction ratio (dimensionless): The factor by which distances shrink under map i; must satisfy 0 ≤ s_i < 1 for convergence.

Worked Example

Problem. Consider the IFS for the Sierpiński triangle defined by three maps, each with matrix A = [[1/2, 0],[0, 1/2]] and translations b_1 = [0,0]^T, b_2 = [1/2, 0]^T, b_3 = [1/4, √3/4]^T. (a) Compute det(A) for each map. (b) Starting from the unit equilateral triangle with area √3/4, find the total area after 3 iterations of the full IFS. (c) Confirm that the fractal attractor has zero area.

det⁡(A)=14,Area3=(34)3 ⁣⋅34=273256,d=log⁡3log⁡2≈1.585\det(A)=\tfrac{1}{4},\quad \text{Area}_3 = \left(\tfrac{3}{4}\right)^3\!\cdot\tfrac{\sqrt{3}}{4}=\tfrac{27\sqrt{3}}{256},\quad d=\tfrac{\log 3}{\log 2}\approx 1.585

Answer. det(A) = 1/4 for each map; after 3 iterations total area = 27√3/256 ≈ 0.182; in the limit the attractor has zero area and fractal dimension log(3)/log(2) ≈ 1.585.

Figure
Diagram
Sierpiński Triangle: 3 IFS Maps & Area ScalingOriginal triangle (area = S)f₁(T)f₂(T)f₃(T)×(3/4)per stepArea→0as n→∞Area Scaling|det(A)| = (½)² = ¼3 copies × ¼ = ¾Total area × ¾ each step
Origins · 1 minWacław Sierpiński and how this was discovered
WS
Wacław Sierpiński
1882–1969 · Polish

Sierpiński introduced the Sierpiński triangle in 1915 and made foundational contributions to set theory, number theory, and topology. His construction of self-similar sets predated the formal theory of fractals by decades and provided concrete examples of sets with counterintuitive properties such as zero area but infinite boundary length. Sierpiński published over 700 papers and 50 books during his career, making him one of the most prolific mathematicians of the 20th century — even though his country was occupied and disrupted by two World Wars.

  • Determinant: Geometric Interpretation — The determinant is the core tool for measuring how a matrix scales area — directly used in every step of the fractal dimension calculation.
  • Matrix Linear Transformations — IFS maps are affine transformations built from matrices; understanding how matrices encode rotations and scalings is essential for constructing fractals.
  • Eigenvalues and Eigenvectors — Eigenvalues of an IFS matrix control contraction along specific directions, linking spectral theory to the geometry of the fractal attractor.
  • Fixed Points and Convergent Iterations — Banach's Fixed-Point Theorem, which guarantees the existence of a unique fractal attractor, is the fundamental convergence result behind all IFS theory.