Designing pedestrian zones within city center networks considering policy objective trade-offs

Abstract

An idea key to human-centric city planning is the reclamation of urban spaces from vehicles and their redesign for human activities. Aligned with this concept, this study presents a framework for human-centric network design. The framework is developed based on a multi-objective optimization model that designs pedestrian zones within a city center network considering conflicting policy objectives. Design performance is evaluated through the interaction between the design and the behavior of network travelers. A vehicle–pedestrian multimodal network equilibrium assignment model is constructed to this end. To efficiently search for better designs, we also develop a hyper-heuristic based on the adaptive large neighborhood search algorithm that dynamically adjusts the probability of selecting operators for neighborhood search. The framework was applied to a city center network in Kawagoe City, Japan, where the conflict between vehicular and tourist pedestrian traffic has long been considered a major problem. Our algorithm successfully found a set of Pareto frontier solutions that clearly show the trade-off between conflicting objectives. A balanced network design among the frontier solutions improved pedestrian comfort by 64.7% while increasing vehicular travel time by only 3.8%. It would be ideal for municipalities to improve the pedestrian experience while maintaining current levels of convenience to drivers, but both issues must be weighed against severe road space constraints. Our approach can aid the discussion on this trade-off by providing a set of Pareto frontier solutions, as each frontier solution shows a different trade-off pattern and can be considered as a meaningful design alternative for policymakers.

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 — 12 citations exist upstream.

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