| Summary: | Efficient routing optimization yields benefits that extend beyond mere financial gains. In this thesis, we present a methodology that utilizes a graph convolutional neural network to facilitate the development of energy-efficient waste collection routes. Our approach focuses on a Waste company in Tromsø, Remiks, and uses real-life datasets, ensuring practicability and ease of implementation. In particular, we extend the dpdp algorithm introduced by Kool et al. (2021) [1] to minimize fuel consumption and devise routes that account for the impact of elevation and real road distance traveled. Our findings shed light on the potential advantages and enhancements these optimized routes can offer Remiks, including improved effectiveness and cost savings. Additionally, we identify key areas for future research and development.
|
|---|