Abstract

Simulating road networks is important for conducting simulation studies of crime occurring on or near roads. In spatial statistics, the concept of homogeneity is well-defined for point patterns, namely constant intensity across an observation area, and simulating homogeneous point patterns can be considered a solved problem. The same cannot be said for more complex structures like line patterns or spatial networks, which are used to model road networks. Recent research has extended the definition of homogeneity to spatial linear networks. This research uses that definition to simulate explicitly homogeneous spatial networks. Specifically, we investigate how spatial networks can be generated from homogeneous point patterns, such that the spatial networks inherit the property of homogeneity from the underlying point patterns. This research thus aims to establish a framework for simulating spatial networks that preserve the statistical properties of homogeneity, ultimately providing a benchmark for analysing real-world networks.