An Intelligent Deep Learning based Classification with Vehicle Routing Technique for municipal solid waste management

Abstract

Municipal solid waste (MSW) management is a critical challenge in urban areas due to increasing waste production and its environmental impact. This study presents an Intelligent Deep Learning-driven Classification with Vehicle Routing (IDLCVR-MSW) system to enhance waste classification accuracy and optimize transportation efficiency. The classification model integrates YOLOv3 for object detection, enhanced with ResNet-50 and XGBoost, achieving a high accuracy of 98.88 %, surpassing existing models such as MobileNetV2 and ResNet-50. To optimize waste collection routes, an Improved Moth Flame Optimizer (IMFO) incorporating Levy flight is implemented, reducing transportation costs by 15–20 % and greenhouse gas (GHG) emissions by 12–18 % compared to traditional methods like Particle Swarm Optimization (PSO). Experimental validation on real-world datasets confirms the model's effectiveness in improving operational efficiency and sustainability. The proposed system supports smart city initiatives by reducing waste collection costs, minimizing environmental impact, and promoting efficient resource utilization. Future work should explore IoT-enabled smart bins and renewable-energy-based waste collection vehicles to further enhance waste management strategies.