Tom Eckhardt
Research Subjects
Research Interests
Tom Eckhardt has a total of 3 co-authors. Here are the top co-authors by number of shared publications:
- René Unger
EA Systems Dresden GmbH, 01187 Dresden, Germany
1 shared publication - Tobias Rodemann
Honda Research Institute Europe, Offenbach, Germany
1 shared publication - Torsten Schwan
EA Systems Dresden GmbH, 01187 Dresden, Germany
1 shared publication
Latest Publications
- Open Access
Using Agent-Based Customer Modeling for the Evaluation of EV Charging Systems
The development of efficient electric vehicle (EV) charging infrastructure requires a modeling of customer behavior at an appropriate level of detail. Since only limited information about real customers is available, most simulation approaches employ a stochastic approach by combining known or estimated customer features with random variations. A typical example is to model EV charging customers by an arrival and a targeted departure time, plus the requested amount of energy or increased state of charge (SoC), where values are drawn from normal (Gaussian) distributions with mean and variance values derived from user studies of obviously limited sample size. In this work, we compare this basic approach with a more detailed customer model employing a multi-agent simulation (MAS) framework in order to investigate how a customer behavior that responds to external factors (like weather) or historical data (like satisfaction in past charging sessions) impacts the essential key performance indicators of the charging system. Our findings show that small changes in the way customers are modeled can lead to quantitative and qualitative differences in the simulated performance of EV charging systems.
Open AccessUsing Agent-Based Customer Modeling for the Evaluation of EV Charging Systems
The development of efficient electric vehicle (EV) charging infrastructure requires a modeling of customer behavior at an appropriate level of detail. Since only limited information about real customers is available, most simulation approaches employ a stochastic approach by combining known or estimated customer features with random variations. A typical example is to model EV charging customers by an arrival and a targeted departure time, plus the requested amount of energy or increased state of charge (SoC), where values are drawn from normal (Gaussian) distributions with mean and variance values derived from user studies of obviously limited sample size. In this work, we compare this basic approach with a more detailed customer model employing a multi-agent simulation (MAS) framework in order to investigate how a customer behavior that responds to external factors (like weather) or historical data (like satisfaction in past charging sessions) impacts the essential key performance indicators of the charging system. Our findings show that small changes in the way customers are modeled can lead to quantitative and qualitative differences in the simulated performance of EV charging systems.