车辆实际工况是一个随机动态过程，其概率分布由道路环境、驾驶员意图、车辆状态等多种因素共同决定。传统的混合动力系统能量管理策略用确定性模型来研究这一随机过程，往往只能得到瞬时最优解或只适用于特定工况的全局最优解，在实际应用中受到极大限制。本研究针对车辆行驶过程的随机性特点，采用统计决策与优化的方法，将混合动力系统能量管理模拟为一个以电池为能量仓库，电能存取服从实车工况统计分布的随机存储模型，根据在线获取的车辆行驶工况统计特征，以系统获得最大节能期望值为目标进行能量的分配与管理。首先对并联式混合动力系统多模式能量流动关系进行定量分析，提取出各节能模式及节能效果定量评价方法，为统计优化提供价值函数；然后结合工况统计特征确定能使节能模式最大化实现的SOC控制目标，并根据随机存储优化理论获取节能期望值最大的模式切换与功率分配规律。最终形成一套能自适应工况变化、且能实现节能期望值最大化的能量管理策略。

The real world driving conditions of vehicle is a stochastic dynamic process, the probability distribution of which was decided by the road environment, the intention of the driver, the vehicle status and other factors. Traditional energy management strategies of hybrid systems using a deterministic model to study this random process, which can only get an instantaneous optimal solution or globally optimal solution only applicable to a specific condition. These strategies were limited in practical applications. Considering the random characteristics of vehicle driving conditions, a statistical based decision-making and optimization method was used in this research. Energy management of hybrid system was simulated as a random inventory model which use battery as an energy storehouse, power access obey the real-turner conditions of statistical distribution of driving parameters. Energy distribution and management strategies were optimized to achieve the target of maximum expectations of energy-saving, based on online access of the statistical characteristics of vehicle driving conditions. Firstly, a quantitative analysis of energy flow relationship between multi-mode of parallel hybrid system was carried out, energy-saving modes and evaluation method of energy saving effects were extracted, which can be used as the value function for statistical optimization; Then, control objectives of SOC to achieve maximum energy-saving mode was determined by considering the statistical characteristics of driving parameters. Based on which, mode switching and power distribution law to achieve maximum expectation of energy saving were established using random storage optimization theory. Eventually, an energy management strategy which can adaptive the change of working conditions and can achieve maximum expectation of energy saving was established.