Topological Clustering for Geographical Networks - Methods for Multilevel Analysis and Visualisation of Geographical Networks - page 49

maintenance of at least one path between these nodes when some network
components happen to be unavailable.
These criteria are exclusively topological because they are based on the presence
of edges, the connection quality or the connection robustness between node pairs.
Nevertheless, graph theory provides extensions that integrate the edge directions and
values, and geography itself provides suitable models that help to take the specific
features of infrastructure networks into account.
From then on, the clustering criteria become relevant within an application
context because they provide a topological framework within which network
simplifying methods can be contextually adapted. However, this approach leaves
several issues unaddressed. These issues need to be considered based on the
meaning of the topological clustering within the context of the geographical study:
How can cluster quality be assessed? How can an index be computed for a simplified
network? How can nodes belonging to several clusters be managed? How is a path
on a network affected by its simplification? Is it possible to simplify a network that
has already been simplified?
Fig. 3.18 Clustering on road networks for simplification of junctions from Mackechnie and
Mackaness ( 1999 )