PHEV(Plug-in hybrid electric vehicles) is a research hotspot in developing new energy vehicles and alleviating pollution pressure, but its cost feasibility and charging planning are two major problems. Past research has been mostly about improving the mismatch of electric vehicles charging demand and infrastructure or assuming PHEV can charge everywhere. This neglects the inherent demand balance of PHEV between refueling and charging and under-utilized charging stations. Based on data mining technology, this research uses trajectory data of 1000 taxis in Shanghai in a month to simulate the actual travel patterns of PHEV. We then estimate the ideal electrification rate and the actual electrification rate of PHEV considering the drive switch between gasoline and electricity and establishes the the optimal layout model solving by the immune algorithm and DBSCAN algorithm which choosing the electric rate and the sum of travel distance as the indicators. Although this study cannot totally represent the real-world PHEV travel patterns, it can still improve our understanding of PHEV’s characteristics and help make better decision . The results indicate that (1)the optimal layout of charging stations match the maximum electrification rate and the minimum sum of travel distance from demand points to charging stations;(2)different battery sizes greatly influence the optimal layout and bigger battery size brings higher electrification rate; and(3)the real electrification rate must be less than the ideal electrification rate using the same parameters. Keywords: charging station; layout optimization; PHEV; trajectory data; electrification rate;

CICTP