Electric load forecasting: A multi-agent systems approach

While electrification of the transportation system will help reduce greenhouse gas emissions and our dependency on imported oil, the growing number of Plug-in Hybrid Electric Vehicles (PHEVs) can also impose a significant new load on electric grids. In this paper, we develop an agent-based behavioral transportation model of PHEVs for the purpose of electric load forecasting in large urban areas. This model is intended to simulate the temporospatial dynamics of PHEVs' behavior due to the changes in a multitude of internal PHEV-related transportation attributes. The results from simulations highlighted the inverse correlation between PHEVs' charging time and peak electric load consumption. The proposed model did not attempt to fully explain the very complex behavior of PHEV agents within the transportation infrastructure, but was rather developed to provide a theoretical foundation to investigate any emergent temporospatial outcomes. The developed model is capable of incorporating external attributes such as collisions, and weather conditions for any future research.