This study describes a method for programming a plug-in electric vehicle (PEV) agent that can be used in power system models and in embedded systems implemented in real PEVs. Implementing the software in real-life applications and in simulation tools enables research with a high degree of detail and practical relevance. Agent-based programming, therefore, is an important tool for investigating the future power system. To demonstrate the PEV agent behavior, an optimization algorithm is presented and two battery aging methods, as well as their effect on vehicle-to-grid operation, are analyzed. Aging costs based on the depth-of-discharge result in shallow cycles and a strong dependency on driving behavior, because the state-of-charge affects the discharging process. In contrast, aging costs based on energy throughput calculations result in deeper cycles and V2G operation, which is less dependent on driving behavior.
