The survivability of cycling in a co-evolutionary agent-based model

Abstract

Many cities plan to grow cycling as a prominent mode to improve accessibility and environmental and financial sustainability. However, relatively few cities have made meaningful headway in this direction. Policymakers would be more inclined to implement the necessary interventions when they have certainty about potential demand, especially knowing where it is located in space. This paper introduces an approach to estimating potential cycling demand using agent-based modelling to determine who may benefit from switching from their current modes to cycling. People benefit when they obtain a similar or higher travel utility score when cycling between their daily activities than when using their existing modes. The model is based on individual mode selection, that all activities in the trip chain are included and can include detailed road and cycle network elements. The co-evolutionary mechanisms within the agent-based simulation allow us to test the potential for cycling relative to the performance of other modes on the network. The case for Cape Town, South Africa, shows that about 32% of those that travel would benefit from cycling based on their utility score. Understanding that travel time benefits are not the only criteria for mode selection, we apply a rejection sampling algorithm based on demographic factors to demonstrate that a more realistic, or pragmatic, cycling potential for Cape Town is in the region of 8%. The results also show that more than 46% of the observed pragmatic demand for cycling is concentrated in an area covering less than 7% of the study area. This has practical implications for policymakers to target interventions both in space and towards specific demographic market segments.