Data-driven synergistic optimization of urban carbon and energy dynamics: Strategies for regional low-carbon...

To address the escalating tension between economic development, energy utilization, and carbon constraints amid rapid urbanization, this study develops a multi-level interactive optimization framework for facilitating regional low-carbon transition. The framework integrates a Back Propagation Neural Network optimized by Particle Swarm Optimization with the Non-dominated Sorting Genetic Algorithm II to simulate policy scenarios and identify coordinated development pathways. Using Guizhou Province as a case study, the optimized results show that, compared with the baseline scenario, carbon emissions are reduced by 6.56 × 10⁷ t, gross domestic product increases by 5.49 × 10¹⁰ CNY, and energy consumption decreases by 3.59 × 10⁷ t. The optimization results reveal distinct strategic orientations across cities. Guiyang and Zunyi emerge as economically pivotal cities under the optimized scenario, whereas Qianxinan and Qiannan exhibitstronger catch-up development needs, and Bijie and Qiandongnan appear better suited to prioritizing low-carbon pathways. By linking multi-objective optimization with inter-city heterogeneity, this study provides a data-driven analytical framework for understanding regional low-carbon transition under the "dual-carbon" goals.