Which urban configurations enhance urban vitality? Nonlinear thresholds and genetic-algorithm optimization for the first-phase planned new towns in South Korea

초록

This study examines the relationship between built environment configurations and nighttime urban vitality in South Korea's first-phase new towns and derives scenario-specific configuration typologies using an optimization approach. Nighttime vitality was measured using the VIIRS Nighttime Lights data as an illumination-based proxy for after-dark activity intensity. A Spatial Generalized Additive Model with two-dimensional spatial smooth estimates non-linear associations between vitality and planning elements. A genetic algorithm then uses the fitted model as a fitness function to search for high-performing configurations under constraints for commercial, medium/high-density, and low-density residential areas. The results revealed three distinct response patterns. Population density and floor area ratio exhibited broadly monotonic positive associations, while the building coverage ratio showed a turning point near 30%, and road density increased more sharply beyond approximately 200 m/ha. Intersection density and mobility-support facilities displayed diminishing returns beyond identifiable thresholds. Across scenarios, the optimized typologies commonly featured higher building coverage, supportive street network conditions, and improved subway accessibility. In peripheral settings, reinforcing local morphology and street connectivity partly compensates for limited transit accessibility. These findings suggest that resilient regeneration planning should emphasize morphological density, street network permeability, and functional diversity, rather than development intensity alone. The cross-sectional design and illumination-based measurements may limit causal inference and underrepresent residential activity.

키워드