Predicting Desire Paths: Agent-Based Simulation for Neighbourhood Route Planning

Abstract

This study addresses how informal “desire paths” emerge when urban infrastructure does not align with pedestrian preferences, focusing on small, local neighbourhoods to bridge a recognised research gap in urban planning. It introduces an agent-based modelling (ABM) framework that builds on path optimisation, affordance theory, and two novel models: a weighted preferences model that accounts for varied pedestrian priorities beyond route length and an obstacle avoidance model that achieves 60 % accuracy, indicating it can successfully replicate aspects of natural navigation, but still showing potential for further refinements. Conducted in three-dimensional simulations using Unity3D and the Agent-Based Modelling for Unity (ABMU) toolkit, this approach yields complex path networks that illuminate pedestrian decision-making processes and highlight the impact of diverse priorities, including local optimisation of route length, on network efficiency. By clarifying these ABM-driven insights, we underscore the significance of integrating multiple pedestrian preferences into urban planning to foster more sustainable development in local neighbourhoods.