A waste extended input-output-based Transformer-LSTM method for analyzing hazardous waste reduction patterns: A case study of shanghai

Abstract

Population growth, economic development, and industrialization have resulted in a notable surge in the hazardous waste (HW) generation, consequently leading to adverse effects on human health and environmental safety. In this work, a Waste Extended Input-Output-based Transformer-Long Short-Term Memory (WEIO-TL) method was developed to simulate the HW generation patterns of various economic industries in the supply chain and generate the desired strategies for reducing HW generation. The WEIO-TL method was applied to Shanghai to identify its HW reduction path and potential by 2035. Multiple scenarios were designed to investigate the impacts of different degrees of economic and technological development on HW generation. The primary results indicated that the economic proportion of each key industry and the technological innovation capability were the main factors influencing the HW generation in Shanghai. The six key industries identified by the WEIO were the chemical industry (CI), other services (OS), transportation equipment (MTE), computers, communication, and other electronic equipment (MCCOEE), wholesale and retail trades (WRT), and construction (C). HW generation would peak at 1570.9 kt in 2025 and subsequently decrease to 949.0 kt by 2035 under the optimal scenario based on the optimal TL model (R2 and RMSE values of 0.976 and 5.968), associated with industry structure optimization (i.e., part of high-polluting industries would be shifted to service) and technological innovation capability improvement. The results of the study will serve as a theoretical foundation to guide other cities toward achieving "zero waste."