Abstract
The evolving patterns of pollutant concentrations and their rigorous assessment are critical issues in contemporary environmental research and policy-making, with important practical implications for air quality management and regional pollution control. To better support such decisions, scientifically sound multi-criteria ranking methods have become a key research focus. In this paper, we propose a novel adaptive functional piecewise ordered weighted averaging (FP-OWA) method for ranking complex functional data. The method extends the existing functional piecewise ranking-weighting framework by integrating data smoothing, depth-based centrality measures, and rank-based aggregation. We systematically compare the performance of FP-OWA with several existing functional data ranking methods using Monte Carlo simulations. The results show that FP-OWA substantially improves ranking consistency and stability when the data are contaminated by white noise. We further apply FP-OWA to rank the daily average PM2.5 and O3 concentrations in 13 cities in the Beijing-Tianjin-Hebei region in 2023, accurately revealing the spatiotemporal differentiation patterns of regional pollution. These findings provide a solid technical basis for local governments to design pollution control strategies and improve air quality. Future research will focus on extending FP-OWA to highly nonlinear and complex functional data, further enhancing its computational efficiency to meet big-data processing requirements, and exploring additional application scenarios.