Load-scheduling and PHEVs in the smart grid

Load peaks can have a negative impact on the stability of the power grid and maintenance costs for transmission and generation companies. Currently, increasing use of plug-in hybrid electric vehicles (PHEV) further proliferates the problem because charging patterns are expected to coincide with peak demand hours, especially the afternoon peak hours. To ameliorate this problem, we have investigated two different multi-agent mechanisms for scheduling the charging PHEVs: one where the charging plans for the PHEVs are created in a centralized manner, and a decentralized mechanism where PHEV agents stochastically generate charging plans on their own accord based on a predefined probability distribution. We found that both mechanisms were able to schedule the PHEVs so that aggravating peak load was avoided, but that it came at a cost of the PHEVs' capacity to recharge in the centralized version. The decentralized mechanism, on the other hand, showed promising results in terms of fairness to the end user while being able to charge PHEVs outside of peak hours.