Abstract: The emergence of electric vehicle-hailing has not only made citizens’ transportation modes more flexible, but also satisfied the needs of personalized and high quality travel. But the way of rational distribution of charge station is a key on account of the battery characteristics of electric vehicle. Thus, this paper founds a basis of the shortest travel time, makes a multi-agent simulation with Python, considers charging behavioral decision problems of internet-based electric vehicle drivers, builds a behavioral model of electric vehicle. What is more, this paper takes 13 charge stations of Beijing as an example, analyzes the rationality of charge stations location planning and makes an optimization of charge stations’ location planning and charging piles’ numbers for a relative balance of access frequencies of every charge station from the perspectives of access frequencies, detour time, waiting time and so forth.

Key words: electric vehicle, charging stations location, multi-agent

摘要： 电动网约汽车的出现使得人们的交通出行方式更加灵活，也更大程度满足出行个性化和高品质的需求，但是由于电动汽车电池特性，如何合理布局充电站成为关键。因此在建立总出行时间最短模型基础上，通过Python对充电站布局进行多Agent仿真优化，平衡交通密度和充电需求，考虑电动网约车司机充电行为决策问题，构建电动汽车行为模型，并以北京市13座充电站为例，分别从访问频率、绕行时间和等待时间等因素分析充电站布局的合理性，对充电站布局以及充电桩个数进行优化，保证各个充电站访问频率较平均。

关键词: 电动汽车, 充电站布局, 多智能体