A comparative analysis of forecasting algorithms for predicting municipal solid waste generation in Chittagong City

Abstract Implementing sustainable solid waste management strategies depends on accurately predicting municipal solid waste (MSW). This study forecasts Chittagong City's waste production using the well-known Multilayer Perceptron (MLP), Support Vector Machine (SVM), and Gaussian Algorithm (GA). The model performance is evaluated based on Mean Absolute Percentage Error (MAPE), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE). Among these, the MLP algorithm demonstrates the highest accuracy in predicting future MSW generation. Waste compositions such as food, fabric, plastic, paper, and wood are also forecasted. Results indicate that by 2030, Chittagong will generate approximately 2,780 tons per day (TPD) of MSW, requiring 247.5 m 2 of landfill space and emitting 51,183.57 tons of greenhouse gases (GHG) under the current waste management practices. This forecast supports decision-makers in modifying and updating waste management systems to achieve sustainability goals, highlighting the practical benefits of accurate predictions in resource optimization, environmental impact mitigation, and long-term planning.