Using shortest path routing to assess cycling networks

Abstract

The lack of cycling infrastructure is a major deterrent to cycling for transportation in the U.S., yet planners and engineers may lack the tools to assess and communicate the potential network impacts of proposed cycling infrastructure. Without these tools, cycling infrastructure may be built ad hoc or where it is politically convenient, instead of where it would be most effective at improving mobility, accessibility, and the actual and perceived safety of cycling. To address these problems, this research outlines a framework for assessing new and existing infrastructure through BikewaySim. BikewaySim is an open-source shortest path model that uses link impedance functions to account for cyclists' preferences for link attributes (e.g., presence of a bike lane, flat terrain, traffic volumes, traffic speeds) and Dijkstra's algorithm to find the least impedance path between any origin-destination pair. To demonstrate the framework, the shortest path is found between 28,392 potential origin-destination pairs for a small study area in Atlanta, GA to assess the impacts of two infrastructure projects. Two impedance functions are demonstrated. The first is based on travel time, while the second uses travel time and example attribute impedances. From the solved shortest paths, percent detour, change in impedance, and link betweenness centrality are calculated, and bikesheds, selected routes, and link betweenness centrality are visualized. Results demonstrate that BikewaySim can effectively visualize potential network improvements from cycling infrastructure and has additional applications for trip planning. Future research will focus on calibrating link impedance functions with GPS trace data showing revealed preferences and surveyed response data on user preferences for cycling infrastructure.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 10 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.