Computational optimization approaches for distributing service points for mobility applications and smart charging of electric vehicles

Abstract

For many business models in the mobility domain an optimal distribution of service points in a customer community is needed. Examples are charging stations of electric vehicles (EVs), bicycle sharing stations, battery swapping stations, or repair stations. Two main challenges are to get the necessary data about the community and environment in order to estimate user demands, local constraints of potential locations, and other properties and to identify optimal service station locations based on these data. Traditionally, these two tasks are considered in a separated fashion. Obtaining input data for the optimization step in a classical way essentially always is inherently incomplete and error prone for larger practical scenarios since manifold aspects play roles in complex, often non-obvious ways, and not all of them can be captured with appropriate estimations of their impacts. In this thesis three different problems are investigated. As an introduction to distributing service points for mobility applications, the first problem focuses on the challenge of identifying optimal locations for battery swapping stations of electric scooters under the assumption that demand information is already known. The second problem is more generic but therefore considers user interaction for obtaining demand information interleaved with the optimization. Instead of estimating customer demands upfront, customers are incorporated directly into the optimization process, i.e., users can interact with the optimization algorithm by expressing their preferences for where to best place service points. Finally we consider the problem of scheduling the charging of EVs at a single charging station in which the maximum power at which a vehicle can be charged depends on the current state-of-charge of the vehicle.