Falconer's Fractal Geometry: Mathematical Foundations and Applications (third edition 2014, first edition 1997) is a fairly traditional text, aimed at higher undergraduate or graduate mathematics students and not so accessible to those from other disciplines. Falconer makes some concessions: he keeps the measure theory to a minimum and states the occasional theorem without giving a proof, while sections with harder material are marked with asterisks and "may be omitted without interrupting the development". And a few sections are almost entirely descriptive, for example the brief look at the Lorentz attractor in "Continuous Dynamical Systems". But otherwise the presentation is rigorous, and clear but dense.
It will help readers if they have some familiarity with fractals already. An introductory chapter provides only a tiny bit of context, rapidly introducing classic fractals such as the Cantor set and the Sierpinski gasket, more for later reference than as motivation. After that Falconer gets straight down to the nitty-gritty.
Part one, "Foundations", begins with a chapter of assumed mathematics: metric spaces and suchlike, measures and mass distributions, a bit of probability. It then proceeds with chapters on box-counting dimension, Hausdorff and packing dimensions, methods for calculating dimensions, the extension to fractals of concepts such as density and tangents, and constraints on the dimensions of the projections, products and intersections of fractals. There's no attempt to be comprehensive here, but rather to introduce the key concepts, prove some major theorems and results, and give a feel for how different concepts of dimension work and are used.
Part two, "Applications and Examples", is broader in scope, but the applications are largely to other areas of mathematics and the examples mostly pull out individual interesting bits of mathematics. The topics covered include iterated function systems (with a glance at image compression), fractals in number theory (in digit distributions, and in continued fractions and Diophantine approximation), dimensions of graphs (self-affine functions and coastlines), Julia sets and the Mandelbrot set, and random fractals; a brief glance at dynamical systems does no more than touch on the logistic map. There is a final chapter at the end on "physical applications", but even this mostly highlights bits of associated mathematics: a continuous model for diffusion-limited aggregation, constraints on the dimension of singularity sets of potential functions, and single aspects of turbulence, antennas, and finance.
So Fractal Geometry is rewarding, but to those who can appreciate mathematics for its own sake. There are exercises at the end of each chapter, with full solutions available online.
Barnsley's Fractals Everywhere (third edition 2012, first edition 1998) takes quite a different approach, building on the core concept of an iterated function system and not introducing concepts of dimension until nearly half-way through. The approach is formal and rigorous (though a few proofs are omitted) and in many ways quite abstract, but the coherent focus on iterated function systems and a plethora of concrete examples, especially ones with two-dimensional visualisations, provide good intuition and motivation.
Barnsley begins with basic metric spaces, then introduces the space of fractals (defined very broadly as the space of compact subsets of an underlying space, with the Hausdorff distance as a metric) and the key notion of an iterated function system (broadly, a complete metric space with a set of contraction mappings) and proves key Contraction Mapping and Collage theorems. He then introduces "code space" and symbolic addressing and dynamics on fractals, working up to the Shadowing Theorem showing that inaccurately calculated orbits are still useful. There's also a brief overview of different approaches to fractal dimension.
The second half of Fractals Everywhere applies and generalises this toolkit. So the construction of fractal interpolation functions is done using iterated function systems, Julia sets are presented as the attractors of iterated function systems, measure theory allows a formal definition of a fractal as a fixed point in a space of measures, and the final chapter uses recurrent iterated function systems as a tool for designing fractals. And there are other goodies, such as a generalisation of the Mandelbrot set to the notion of a parameter space.
There's a good selection of exercises in Fractals Everywhere, cleverly mixed up with examples, so one finds oneself attempting them almost before realising it; there are also detailed solutions for most of them. The black and white illustrations and diagrams, though relatively low resolution, are effective. There are also thirty two pages of colour plates, half of them supporting the main text, mostly illustrating applications to art and image processing, and half forming a kind of photo-essay on fractal art.
There are some suggestions for computer exploration, in a few places with some (now old-fashioned BASIC) code. And the applications discussed are mostly to image processing, with little on the physics and engineering aspects of fractals (there is nothing on the Lorentz Attractor, for example). So Fractals Everywhere may particularly appeal to those with a computer science background. It's a nicely developed exposition, however, covering some fascinating mathematics with verve, so I recommend it to anyone with a background in formal mathematics who is curious about fractals.
In A Tale of Two Fractals Kirillov attempts to reach from fairly elementary undergraduate mathematics — assuming only basic analysis, linear algebra, geometry, group theory and so forth — to novel research results and questions. He manages this by focusing on just two fractals, the Sierpinksi and Apollonian gaskets, and a few topics on those: the Laplace operator and harmonic functions on the Sierpinski gasket and Descartes' Theorem and some geometry and group theory on the Apollonian gaskets. So anyone after a general introduction to fractals should look elsewhere.
The approach is fairly rapid and mostly rigorous, but makes no attempt to be systematic. There are a few "in-line" exercises and some more difficult problems. And there's a good selection of diagrams, some of them in colour. More than a quarter of the text is incorporated into "Info" sections providing background which is necessary but peripheral to the main discussion. These are labelled alphabetically, confusingly out of sequence with the rest of the (numerically ordered) material.
Kirillov does a decent job of selecting material with little in the way of prerequisites, but I found many places where things stated as "obvious" were not so clear, and there are places where the terminology or assumptions seem closer to graduate level. There are also a few significant typographical errors, notably in the diagrams, where the labelling often seems inconsistent with the text. Better proofing would have helped, perhaps with more feedback from undergraduate readers.
Despite these problems, A Tale of Two Fractals might appeal to hardy undergraduates who have research ambitions already and want something out of the syllabus to explore. It offers a good variety of material, incorporating a mix of geometry, analysis, algebra, and number theory.
- External links:
- buy from Amazon.com or Amazon.co.uk
- buy from Amazon.com or Amazon.co.uk
A Tale of Two Fractals
- buy from Amazon.com or Amazon.co.uk
- share this review on Facebook or Twitter
- Related reviews:
- Peitgen et al. - Chaos and Fractals: New Frontiers of Science
- Kenneth Falconer - Fractals: A Very Short Introduction
- Manfred Schroeder - Fractals, Chaos, Power Laws: Minutes from an Infinite Paradise
- books about mathematics
- books published by Dover