基于多Agent建模和NetLogo仿真的异质性驾驶行为对交通流的影响

Influence of Heterogeneous Driving Behavior on Traffic Flow Based on Multi-Agent Modelling and NetLogo Simulation

为弥补现有交通流仿真模型中对异质性驾驶行为考虑不足,基于多智能体建模和NetLogo仿真方法研究了异质性驾驶行为对交通流的动态影响。首先,基于规则和模糊控制算法,构建了人-车-路多智能体系统模型。然后,以西安市某路段为案例,构建了3车道城市主干路交通流仿真模型,运用NetLogo软件实现了车辆跟驰、换道、碰撞等行为可视化仿真。最后,分析对比了异质性驾驶人车速控制和换道行为差异,研究了碰撞事件影响下的动态交通流演化特征,并与其他仿真模型进行对比。结果表明:交通事件会影响交通流的连续性,在固定仿真时间条件下,即使碰撞事件消失,拥堵状态缓解,交通流的整体车速只能恢复至最高车速的70%;异质性驾驶人在车速控制和换道行为上存在差异,激进型驾驶人的车速控制策略更加离散。与VISSIM仿真模型相比,基于多智能体的交通流仿真模型能够揭示交通瓶颈和碰撞事件之间的关系,为中微观交通流仿真提供了参考。

To compensate for the inadequate consideration of heterogeneous driving behaviors in existing traffic flow simulation models,the dynamic influence of heterogeneous driving behaviors on traffic flow was studied using multi-agent modelling and NetLogo simulation method.Firstly,a human-vehicleroad multi-agent system model was constructed based on rules and fuzzy control algorithms.Then,taking a road section in Xi′an as a case study,a simulation model of traffic flow on a three lane urban main road was conducted,utilizing the NetLogo software to achieve visual simulation of vehicle following,lane changing,collision,and other behaviors.Finally,the differences in speed control and lane changing behavior among heterogeneous drivers were analyzed and compared,and the dynamic traffic flow evolution characteristics under the influence of collision events were investigated.The constructed model was compared and evaluated with other simulation models.The results reveal that traffic events can disrupt the continuity of traffic flow.Even if the collision events disappear and the congestion is relieved,the overall speed of the traffic flow can only be restored to 70%of the maximum speed at fixed simulation intervals.Additionally,heterogeneous drivers′speed control and lane changing control behaviors vary to different degrees due to traffic dynamics events,with aggressive drivers exhibiting more discrete speed control strategies.Compared wih VISSIM model,multi-agent-based traffic flow simulation model can effectively reveal the relationship between traffic bottlenecks and collision events,providing methodological references for micro-macro traffic flow simulation.