In this paper, we propose an optimization model to select the charging locations of an automated electric taxi (AET) system. The service provided by this AET system is a seamless door-to-door service connected to the train station, which helps improve the last mile transport. We individualize the vehicles instead of treating them as a flow to track the remaining battery level of each AET. Two types of charging are considered containing depot charging with lower charging speed and opportunity charging with higher charging speed. We formulate a mixed-integer programming model with linear constraints to optimize the locations of depot charging and opportunity charging according to the objective function of maximizing the number of satisfied requests. The proposed model is applied to the case study city of Delft, the Netherlands with the travel demand generated by the Delft Zuid train station. Results show that the charging scheme with two types of charging can provide sufficient electrical energy for shared use AETs to serve passengers last mile travel demand.

The design of hybrid electric vehicles based on real-world driving conditions becomes increasingly important due to strict legislation and complexity of future powertrain concepts. Therefore, component sizes should be optimized not only with respect to fuel consumption but also regarding real-world load conditions. Since e.g. a proper design of the cooling system of electrical components requires consideration of time-related power demand, this paper proposes a novel concept of incorporating time frame-based load analysis (TFBA) into driving cycle synthesis. The synthesis is carried out within a two-layer optimization framework where the first layer considers statistical features and the second considers quantities directly related to a vehicles' load implication. By taking into account Markov chain theory, a class frame is derived which then serves as design space for optimizing a sequence of micro-trips with respect to meeting target criteria based on the concepts of mean tractive force and TFBA. To prove practicality, an exemplary driving cycle is synthesized and the method is validated within a parameter study. Results show that the synthetic driving cycle is representative with respect to the considered target criteria and moreover statistical quantities used for validation are within an error tolerance of 5%.