The novel approximated global optimal energy management strategy has been proposed for hybrid powertrains.
Eight typical driving behaviors have been classified with K-means to deal with the multiplicative traffic conditions.
The stochastic driver models of different driving behaviors were established based on the Markov chains.
ECMS was used to modify the SMPC-based energy management strategy to improve its fuel economy.
The approximated global optimal energy management strategy for plug-in hybrid electric buses has been verified and analyzed.