Harmonizing stakeholder interests in urban renewal: A novel planning approach using explainable machine learning and spatial optimization

Abstract

Urban renewal is essential for revitalizing existing urban land and promoting sustainable urban development, with urban renewal planning optimal being a crucial research challenge. Current planning methods neglect the complex interactions and conflicts among multiple stakeholders, and often lack explainability in the site selection process. To address these issues, proposes a Multi-Stakeholder Perspective Urban Renewal Planning Optimization (MSP-URPO) model. This innovative approach integrates a multi-objective optimization algorithm with eXplainable Machine Learning (XML) to enhance decision-making for urban renewal. The optimization algorithm resolves conflicts among multiple objectives, while XML improves the clarity and understanding of the planning results. Applied to a case study in Shenzhen, the MSP-URPO model predicts an urban renewal scale of 616 ha by 2024, selecting 786 optimal blocks from 23,957 candidates. The study reveals that residents' preferences and multi-stakeholder decision consistency significantly impact site selection, contributing 33.69 % and 27.66 % respectively. These findings demonstrate that the proposed method effectively provides a low-cost, efficient, and precise decision-support tool for urban management and renewal planning.