Title: Solid waste generation forecasts using long short-term memory approach

Authors: Aya Idrissi; Rajaa Benabbou; Jamal Benhra; Mounia El Haji

Addresses: Optimization of Industrials and Logistics Systems Team OSIL, Laboratory of Advanced Research in Industrial and Logistic Engineering (LARILE), Department of Industrial and Logistic Engineering GIL, National High School of Electricity and Mechanical Engineering ENSEM, Hassan II University of Casablanca UH2C, Casablanca, Morocco ' Optimization of Industrials and Logistics Systems Team OSIL, Laboratory of Advanced Research in Industrial and Logistic Engineering (LARILE), Department of Industrial and Logistic Engineering GIL, National High School of Electricity and Mechanical Engineering ENSEM, Hassan II University of Casablanca UH2C, Casablanca, Morocco ' Optimization of Industrials and Logistics Systems Team OSIL, Laboratory of Advanced Research in Industrial and Logistic Engineering (LARILE), Department of Industrial and Logistic Engineering GIL, National High School of Electricity and Mechanical Engineering ENSEM, Hassan II University of Casablanca UH2C, Casablanca, Morocco ' Optimization of Industrials and Logistics Systems Team OSIL, Laboratory of Advanced Research in Industrial and Logistic Engineering (LARILE), Department of Industrial and Logistic Engineering GIL, National High School of Electricity and Mechanical Engineering ENSEM, Hassan II University of Casablanca UH2C, Casablanca, Morocco

Abstract: In recent years, Artificial Intelligence (AI) has become increasingly prominent across various domains with the objective of emulating human intelligence, and no industry has escaped its powerful advancement, including the waste management sector. This study proposes a time-series forecasting method based on a Long Short-Term Memory (LSTM) network to accurately predict the amount of waste daily collected. The LSTM model was trained using data points that align with the same phase of the seasonal cycle as the forecasted point, this approach preserves the chronological order of the data while focusing on points corresponding to the same phase of the seasonal cycle. To better illustrate the LSTM neural network's accuracy and robustness, a Back-Propagation Artificial Neural Network (BP-ANN) based on the Levenberg-Marquardt training method was also used. The results demonstrated that the proposed LSTM outperformed in terms of accuracy and precision, proving its excellence in capturing the complex patterns.

Keywords: solid waste generation; artificial intelligence; LSTM; long short-term memory; neural network; back propagation.

DOI: 10.1504/IJAMS.2025.147307

International Journal of Applied Management Science, 2025 Vol.17 No.2, pp.163 - 183

Received: 23 Nov 2023
Accepted: 02 Jun 2024
Published online: 14 Jul 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article