Modeling fractal structure of city-size distributions using correlation functions

Area of Science:

• Urban studies
• Complex systems
• Statistical physics

Background:

• Zipf's law describes city rank-size distribution but lacks a clear explanation for its scaling exponent.
• Existing models do not fully account for the observed range of scaling exponents in urban systems.

Purpose of the Study:

• To propose a dual competition hypothesis explaining the scaling exponent in Zipf's law for urban development.
• To reconcile Zipf's law and Pareto's law within a unified framework of urban evolution.
• To derive the parameter intervals for scaling exponents using fractal and multifractal concepts.

Main Methods:

• Application of general fractal and scaling ideas to urban development.
• Mathematical transformation between Zipf's and Pareto's distributions.
• Construction of frequency and size correlation functions.
• Scaling analysis and multifractal spectrum analysis.
• Mathematical experiments on hierarchical correlation.

Main Results:

• Derived Pareto exponent intervals: (0.5, 1] from Pareto distribution and [1, 2) from Zipf distribution.
• Identified two key effects in urban evolution: Pareto effect (external complexity/city number) and Zipf effect (internal complexity/city size).
• Demonstrated that the scaling exponent varies between 0.5 and 2 due to the interplay of these effects, approaching 1 at equilibrium.
• Validated the model showing that adherence to Zipf's law implies adherence to scaling laws for frequency and size correlations.

Conclusions:

• The dual competition hypothesis provides a robust explanation for Zipf's law and its scaling exponent in urban systems.
• The framework unifies Pareto and Zipf distributions, representing distinct but related urban evolutionary processes.
• The theory has broader applicability to other inverse power-law distributions across scientific disciplines.