Data-driven network optimization in shared mobility era

Abstract

Shared mobility has emerged to fill gaps in today’s transportation networks. Its applications encompass bikesharing, carsharing, ridesharing, on-demand ride services, last-mile delivery, and other shared-use mobility options. The goals of shared mobility are to enable users or goods to gain short-term on-demand access to transportation modes, reduce traffic congestion, and to promote environmental sustainability. While the concept of shared mobility has the potential to drive transportation systems more efficient, there are new operational challenges. The many users and transportation service providers and their complex connectivity require new network optimization methods to achieve the goals. The large volumes of data captured from transportation networks call for novel data analytics techniques to understand users’ demands and behaviors. One of the key challenges in shared mobility applications is the supply-demand imbalance problem. We propose a novel data-driven methodology to cluster demand points into communities for better resource deployment. A numerical study based on a bike-sharing network is carried out to examine the performance of our solution approach.