A multi-AI approach to predicting municipal solid waste generation and recycling demand in Hong Kong

Accurate prediction of municipal solid waste (MSW) generation is recognized as a critical component in establishing optimized waste management frameworks. Traditional regression and single time-series models often prove inadequate in capturing the nonlinear and multifactorial dynamics of MSW generation. To address these shortcomings, this study integrates advanced AI-driven regression methods (e.g.,MLP-ANN) with time-series models (e.g.,LSTM,ARIMA) to enhance predictive accuracy in the context of Hong Kong. By incorporating diverse socioeconomic variables, our approach markedly outperforms conventional techniques, particularly in forecasting food, plastic, and paper waste. Furthermore, aligned with Hong Kong’s recycling targets, we predict the recycling capacity required for 2024–2035. The results underscore the urgent imperative for immediate, large-scale investments in waste recycle infrastructure, especially in food and plastic waste, to mitigate future landfill saturation.