Multi-objective land use optimization for enhanced urban livability using a NSGA-III-differential evolution algorithm

Abstract

Balancing economic growth, environmental sustainability, and social equity to enhance urban livability, an essential dimension of human-centered development, poses significant challenges in metropolitan regions. Conventional planning paradigms frequently struggle with navigating the intricate, multi-dimensional trade-offs embedded within human-centered land use decision-making. To bridge this gap, this study proposes a novel human-centered planning support framework that uniquely integrates a hybrid Non-dominated Sorting Genetic Algorithm III with Differential Evolution (NSGA-III-DE) to optimize land use allocation for urban livability. The core novelty lies in the holistic integration of three components: (1) a comprehensive structure of six critical livability objectives, including a newly introduced accessibility metric; (2) an innovative dynamic Floor Area Ratio (FAR) mechanism that co-optimizes development intensity with land use type, realistically modeling densification trade-offs; and (3) a spatially explicit approach that identifies transformation hotspots aligned with urban renewal priorities. The model simultaneously optimizes 5507 land parcels in Shenzhen’s Nanshan District. Empirical validation demonstrates the algorithm's effectiveness, achieving significant improvements across objectives, including a 26.1 % reduction in land use transition costs and a 13.3 % decrease in commuting intensity, while enhancing accessibility by 7.1 %. The resulting Pareto-optimal solutions reveal key policy trade-offs, and identified change hotspots correspond closely with official urban renewal plans, underscoring the model's practical relevance. This study contributes a robust, data-driven decision-support tool for sustainable urban transformation, offering actionable insights for livability-oriented planning in rapidly evolving, high-density contexts.