Multi-agent Learning for Energy-Aware Placement of Autonomous Vehicles

Mobility gets an increasing amount of meaning and significance in the modern society. In this paper, we introduce a multi-agent learning application for a multi-agent system in e-mobility. In particular, we propose a geospatial model for free-floating and autonomously driving e-trikes and demonstrate a calculation method of positioning e-trikes on a given area by using different methods of cluster analysis. The solution of the cluster analysis contains cluster centers which represent a positioning for the e-trikes. The solution is then evaluateded by a simulation model with more sophisticated parameters. This research field opens different opportunities of application scenarios, which are discussed in the conclusion.