Route choice through regions by pedestrian agents

Abstract

Simulation models for pedestrian movement are valuable tools to support decision-making processes in urban design. However, existing models of pedestrian behaviour are built on simplistic assumptions regarding people's representation of the urban space and spatial behaviour. In this work, a route-choice algorithm that takes into account regionalisation processes and the hierarchical organisation of geographical elements is adapted for pedestrian movement and incorporated into an agent-based model. The macro-level patterns emerging from two scenarios, one employing an angular-change minimisation algorithm and the other employing the regional algorithm here proposed, are compared for a case study in London, UK. Our routing algorithm led agents to recur to a higher number of street segments, i.e. routes were more diverse among agents. Though validation has not yet been performed, we deem the patterns resulting from the regional algorithm more plausible.