Online adaptive algorithm for scheduling PEV charging

Plug-in electric vehicles (PEVs) have been increasing in recent years and are expected to keep growing even more. PEVs are charged by using the electricity from the power grid, which leads to higher electricity demands. If PEVs are charged randomly without proper management, the charging might occur during peak periods. Therefore, the charging could cause severe problems to the power system. One way to mitigate this problem is for utilities to employ time dependent pricing such as Critical Peak Pricing or Real Time Pricing (RTP). This paper presents a Learning Automata (LA) algorithm for scheduling PEV under RTP. To train the algorithm and for a good initial schedule, it is used offline with the previous available data, then the algorithm is used online for the current data. It is shown that for a large fleet of PEVs, it might take a few days to learn the best schedule. However, even during this period, the results are better than unscheduled charging. The paper shows that with a proper scheduling algorithm, we can minimize the charging cost of a large number of PEVs.