Vehicle Routing Problems (VRPs) play an important role in industry for managing and optimizing operations for fleets of delivery vehicles. Our project aims to incorporate new objectives that pertain to social and environmental sustainability together with classical ones, such as time and resource minimization. One key such aspect would be de-congesting the road network. One natural way of combining these aspects are through multi-objective optimization and recently the field of Pareto Set Learning (PSL) has shown much promise in this regard. We aim to apply this machine learning paradigm and investigate the following 1) Can we use this framework to efficiently solve new types of problems, such as time and capacity constrained VRPs with multiple objectives. 2) Can we generalize our method to large graphs and perform well out-of-distribution. 3) Can we optimize the architecture or learning algorithm to better suit the needs of the project.