Urban pedestrian movement is inherently uncertain, shaped by built form, microclimate, and time-varying crowding. This study develop a quantum-inspired probabilistic framework that models pedestrian presence as a spatial probability field derived from a composite potential integrating static structure (connectivity, crossings, barriers, visibility, points of interest) and dynamic drivers (crowd density, shade/thermal exposure). Applying the method to University Junction, Moratuwa, Sri Lanka, the study discretizes the domain to 5 m cells and 15-minute time bins, estimate factor weights via variational minimization, and solve an eigen-problem to obtain probability maps. The model reproduces diurnal reconfiguration of flows, concentrating midday probabilities in shaded, connected corridors and reducing presence on exposed verges. Link-level validation shows strong agreement with observed shares (Pearson’s r ≈ 0.77 evening; ≈ 0.71–0.77 across periods) and realistic spatial autocorrelation. Scenario tests (temporary barriers, event footprints) re-equilibrate the probability field without retraining, revealing predictable re-routing to substitute corridors. Compared with a Boltzmann-type classical model and a space-syntax predictor, the proposed approach achieves higher fit and better spatial realism by explicitly encoding climate comfort and crowding. The framework yields policy-ready maps that support shade investment, crossing consolidation, and operational crowd management, providing an interpretable and transferable tool for assessing, managing, and allocating pedestrian space under uncertainty.
Deshan et al. (2026) studied this question.