微观交通仿真模型参数两阶段标定优化方法

doi:10.16182/j.issn1004731x.joss.25-FZ0635

摘要/Abstract

摘要：

为解决微观交通仿真模型参数标定方法耗时长的问题，提出了两阶段标定优化方法。第一阶段训练基于神经网络的仿真代理模型，建立模型参数与校核指标的映射关系，结合遗传算法筛选候选参数；第二阶段在获取近似最优参数后，以该参数组为初值，结合真实仿真模型再次执行遗传算法寻优，进一步提升标定精度。实验结果表明：第一阶段获得的基本最优参数组合在保障仿真精度的同时大幅度节约标定时耗，第二阶段在基本最优参数组合的基础上能够快速实现遗传算法收敛。该方法在保持预测精度的同时，参数优化效率可提高79.7%，为微观交通仿真模型参数自动化实现提供基础。

关键词: 微观交通仿真, 模型参数标定, 神经网络, 效能提升, 自动化标定

Abstract:

A two-stage calibration and optimization method is proposed to address the problem that parameter calibration methods for microscopic traffic simulation models are time-consuming. In the first stage, a surrogate model based on neural networks is trained to establish the mapping relationship between model parameters and evaluation indicators, and a genetic algorithm (GA) is combined to screen candidate parameters. In the second stage, after obtaining the approximate optimal parameters, by employing this set of parameters as initial values, a genetic algorithm is re-executed by combining the real simulation model for optimization to further improve calibration accuracy. Experimental results show that the basic optimal parameter combination obtained in the first stage can achieve substantial reductions in calibration time while ensuring simulation accuracy, while the second stage enables quick convergence of GA based on the basic optimal parameter combination. The proposed method improves parameter optimization efficiency by 79.7% without compromising prediction accuracy, providing a foundation for the automation of microscopic traffic simulation model parameters.

Key words: microscopic traffic simulation, model parameter calibration, neural network, efficiency improvement, automated calibration

中图分类号:

TP391.9

刘益嘉,周晨静,潘冬等 . 微观交通仿真模型参数两阶段标定优化方法[J]. 系统仿真学报, 2026, 38(6): 1711-1721.

Liu Yijia,Zhou Chenjing,Pan Dong,et al . Two-stage Calibration and Optimization Method for Microscopic Traffic Simulation Model Parameters Based on Neural Network Surrogate Models[J]. Journal of System Simulation, 2026, 38(6): 1711-1721.

图/表 19

图1

图2

图3

图4

图5

图6

图7

表1

表2

表3

图8

图9

表4

图10

图11

表5

表6

表7

表8

参考文献 21

[1]	Benekohal R R. Procedure for Validation of Microscopic Traffic Flow Simulation Models[J]. Transportation Research Record, 1911(1320): 190-202.
[2]	Bhattacharyya Kinjal, Maitra Bhargab, Boltze Manfred. Calibration of Micro-simulation Model Parameters for Heterogeneous Traffic Using Mode-specific Performance Measure[J]. Transportation Research Record, 2020, 2674(1): 135-147.
[3]	Sebastian Buck H, Mallig Nicolai, Vortisch Peter. Calibrating Vissim to Analyze Delay at Signalized Intersections[J]. Transportation Research Record, 2017, 2615(1): 73-81.
[4]	Cheu Ruey-Long, Jin Xin, Ng Kim-Chwee, et al. Calibration of FRESIM for Singapore Expressway Using Genetic Algorithm[J]. Journal of Transportation Engineering, 1998, 124(6): 526-535.
[5]	Gao Yacong, Zhou Chenjing, Rong Jian, et al. Enhancing Parameter Calibration for Micro-simulation Models: Investigating Improvement Methods[J]. Simulation Modelling Practice and Theory, 2024, 134: 102950.
[6]	单籽跃, 周晨静, 韦昌永, 等. 微观交通仿真模型参数标定发展与优化实证研究[J]. 交通工程, 2022, 22(4): 46-53.
	Shan Ziyue, Zhou Chenjing, Wei Changyong, et al. Empirical Study on Development and Optimization of Microscopic Traffic Simulation Model Parameter Calibration[J]. Journal of Transportation Engineering, 2022, 22(4): 46-53.
[7]	林子赫, 周晨静, 荣建, 等. 基于工程实测的加速度-速度参数阈值分析[J]. 交通信息与安全, 2019, 37(2): 107-113.
	Lin Zihe, Zhou Chenjing, Rong Jian, et al. An Analysis of Acceleration-speed Parameter Threshold Based on Actual Engineering Measurement[J]. Journal of Transport Information and Safety, 2019, 37(2): 107-113.
[8]	周晨静, 荣建, 郭琳科. 微观交通仿真模型参数标定结果取值方法研究[J]. 系统仿真学报, 2019, 31(12): 2802-2809.
	Zhou Chenjing, Rong Jian, Guo Linke. Research on the Value Accessing Method for Calibrating Micro Traffic Simulation Model Parameters[J]. Journal of System Simulation, 2019, 31(12): 2802-2809.
[9]	周晨静, 高亚聪, 荣建. 微观交通仿真模型参数标定方法改善研究[J]. 系统仿真学报, 2020, 32(11): 2112-2120.
	Zhou Chenjing, Gao Yacong, Rong Jian. Research on Improvement of Parameters Calibration Method of Microscopic Traffic Simulation Model[J]. Journal of System Simulation, 2020, 32(11): 2112-2120.
[10]	He C, Jin Y, Ong Y S, et al. A Review of Surrogate-assisted Evolutionary Algorithms[J]. Expert Systems with Applications, 2023, 223: 119875.
[11]	Naing Htet, Cai Wentong, Yu Jinqiang, et al. Fine-grained Trajectory Reconstruction by Microscopic Traffic Simulation with Dynamic Data-driven Evolutionary Optimization[J]. IEEE Transactions on Intelligent Transportation Systems, 2025, 26(2): 1930-1950.
[12]	Xie D, Zhang Y, Li Z, et al. Metamodel-based Calibration of Microscopic Traffic Simulation: A VISSIM Case Study[J]. Transportation Research Part C: Emerging Technologies, 2023, 155: 104264.
[13]	Shi Y, Li S, Zhao J. Traffic Simulation Optimization Considering Driving Styles: A Bayesian Optimization Approach[J]. Communications in Transportation Research, 2025, 5: 100128.
[14]	高亚聪, 周晨静, 荣建. 基于行程时间累积分布曲线的微观交通仿真模型参数标定[J]. 公路交通科技, 2021, 38(4): 121-130.
	Gao Yacong, Zhou Chenjing, Rong Jian. Parameter Calibration of Microscopic Traffic Simulation Model Based on Cumulative Distribution Curve of Travel Time[J]. Journal of Highway and Transportation Research and Development, 2021, 38(4): 121-130.
[15]	Marcello Montanino, Vincenzo Punzo. Making NGSIM Data Usable for Studies on Traffic Flow Theory: A Multistep Method for Vehicle Trajectory Reconstruction[M]. Washington: Transportation Research Board of the National Academies, 2013: JRC76757.
[16]	Wang Shuling, Patwary A U Z, Huang Wei, et al. A General Framework for Combining Traffic Flow Models and Bayesian Network for Traffic Parameters Estimation[J]. Transportation Research Part C: Emerging Technologies, 2022, 139: 103664.
[17]	Li Yong, Zhou Shaoping. A New Instantaneous Wavelet Bicoherence for Local Fault Detection of Rotating Machinery[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 69(1): 135-143.
[18]	Errachid M, Essanhaji A, Messaoudi A. RMVPIA: A New Algorithm for Computing the Lagrange Multivariate Polynomial Interpolation[J]. Numerical Algorithms, 2020, 84(4): 1507-1534.
[19]	周晨静, 荣建, 陈春安. 面向微观交通仿真实验的模型参数敏感性分析方法[J]. 北京工业大学学报, 2016, 42(11): 1728-1733.
	Zhou Chenjing, Rong Jian, Chen Chunan. Parameter Sensitivity Analysis Method for Microscopic Traffic Simulation Experiment[J]. Journal of Beijing University of Technology, 2016, 42(11): 1728-1733.
[20]	Manjunatha P, Vortisch P, Mathew T V. Methodology for the Calibration of VISSIM in Mixed Traffic[C]//Transportation Research Board 92nd Annual Meeting.Washington, D.C.: Transportation Research Board, 2013: 1-11.
[21]	Song Rui, Sun Jian. Calibration of a Micro-traffic Simulation Model with Respect to the Spatial-temporal Evolution of Expressway On-ramp Bottlenecks[J]. Simulation, 2016, 92(6): 535-546.

参数名称	参数解释	数值范围
OB_d/m	车辆在纵向能感知到的前车距离范围(影响响应延迟与跟驰起点)	[150, 350]
N_pr	驾驶员同时感知与判断的前车数量，影响反应保守程度	[1, 5]
AXE_average/s	停车状态下车辆之间的平均期望车头间距	[1, 4]
BX_add	安全距离的加法部分，反映最小跟驰容忍程度	[1, 5]
BX_mult	安全距离乘法项，调整随速度增长的安全距离幅度	[1, 6]
ACC_max/(m/s²)	车辆能实现的最大加速度上限	[-6, -2]
ACC_ac	加速过程中的平滑性调节因子	[-2, -0.5]
GAP_min/m	变道所需的最小安全车间距	[0.5, 5]
ABX	驾驶员对邻车横向距离变化的敏感程度	[0.1, 1]
CO_ac	当前车对前车强行变道时的减速容忍度，反映交通流中的合作性水平	[-5, -1]
D/m	换道过程中保持目标车道稳定间距的下限	[50, 250]
T_dis/s	驾驶员在决定换道后可容忍的最大等待时长	[10,60]

网络结构 (隐藏层节点)	网络层数	测试集 MAPE/%	备注
64	1	10.40	简单结构，拟合精度不足
64-32	2	9.08	最终选择，结构紧凑、精度较优
128-64-32	3	8.60	精度最优，但结构复杂、训练耗时
128-64	2	9.10	相比最终方案结构更复杂

参数名称	取值
OB_d/m	251.25
N_pr	3
AXE_average/s	2.77
BX_add	3.54
BX_mult	4.03
ACC_max/(m/s²)	-2.8
ACC_ac	-1.4
GAP_min/m	3.3
ABX	0.64
CO_ac	-2.5
D/m	150.69
T_dis/s	60

参数名称	取值
OB_d/m	250.25
N_pr	3
AXE_average/s	2.59
BX_add	3.57
BX_mult	4.43
ACC_max/(m/s²)	-2.7
ACC_ac	-1.36
GAP_min/m	3.05
ABX	0.60
CO_ac	-2.5
D/m	160.69
T_dis/s	60

车辆类型	一阶段	两阶段优化
汽车	8.61	6.6
货车	10.01	7.9