Recommender Systems for Multimodal Transportation in Smart Sustainable Cities

Abstract

Transportation recommendation systems (RS)s have garnered significant attention owing to their ongoing potential for enhancement. One of the key innovations in this domain is multimodal transportation RSs, which suggest travel routes using a combination of different transportation modes. In this paper, a multimodal transportation RS is introduced, which recommends optimized trajectories based on user preferences. The system involves two main steps, trajectory generation and ranking. In the first step, Particle Swarm Optimization (PSO) is used to find optimal trajectory combinations between the origin and destination, followed by post-processing. In the second step, the generated trajectory is evaluated using a RankNet model trained on historical user data with a content-based approach. The results demonstrate the system’s ability to generate feasible trajectories and provide precise recommendations. The results enable an efficient usage and convenient user experiences and may foster the broader use of public transportation combined with other transport modes addressing the objectives of smart and sustainable future cities.