Generating urban-like networks with DLA

Ignas Krikštaponis, Xiang Li, Sebastiaan Kruize, Youri Moll

Getting started

To install all necessary packages, run:

  pip3 install requirements.txt

Usage examples

Notebook with usage examples of the code can be found in examples.ipynb

Modules

• dla_model_final.py - functions for generating DLA arrays (including dynamic stickiness), statistic visualisations and calculating fractal dimension
• networks.py - functions for generating networks from DLA, retrieving real-life networks, converting city networks into fractals, visualising and performing analysis
• animate_grid.py - a function for generating animated gifs from DLA numpy arrays
• fractal_city.py - code for retrieving fractals from real-life city networks.

Other

• experiments.py - code used for running experiments with changing stickiness and number of walkers; and saving the results into .npy files
• result_analysis.py - code used for reading .npy DLA arrays in results folder and extracting network statistics from them
• analyse_cities.py - code used for gathering network statistics for selected real-world cities.

Results

Results generated with experiments can be found in the results folder:

• changing_stickness_250_walkers DLA numpy arrays generated with experiments.py changing stickiness with 250 walkers. file format x,y.npy: x - stickiness, y - simulation number
• changing_stickness_400_walkers DLA numpy arrays generated with experiments.py changing stickiness with 400 walkers. file format x,y.npy: x - stickiness, y - simulation number
• changing_walkers DLA numpy arrays generated with experiments.py changing number of walkers with stickiness of 1. file format x,y.npy: x - number of walkers, y - simulation number
• real_cities - csv files with network statistics for selected real-world cities generated with analyse_cities.py
• periods - a number of DLA .npy arrays generated with periodic stickiness.