-
摘要: 为带延迟过境O-D流估计问题建立了概率分布约束的多元线性回归模型, 设计了求解算法, 利用这个模型分别估计出驶入车流量、过境车流量占路段车流量的百分比和流出车流量的大小, 并且对模型进行了样本数据有误差时的抗差分析。结果发现当路段交通量数据有不超过3%的误差以及路段上不同去向交通流量的比例有较小摆动时, 该模型总体结果能达到8%的相对误差精度, 具有较好的抗干扰能力, 是可行的。Abstract: This paper gave a multi-linear recursive model with probability-constraints, and designed an algorithm following the viewpoint of the model for estimating the O-D matrix through a region with lagged traffics. The stability of the model was analyzed in the conditions of experimentation errors. The estimation error of the model is about 8%, when the error of traffic flow is less 3%, which shows that the model is feasible.
-
Key words:
- traffic planning /
- O-D matrix /
- recursive estimation /
- flow through a region /
- lagged traffics
-
表 1 计算精度与流量结构的关系(1)
Table 1. Simulation result of model (5)
序号 RATE (H) ARE (X) ARE (Y) ARE (H) 1 0.1846 0.0237 0.0254 0.0712 2 0.2498 0.0289 0.0295 0.0501 3 0.3111 0.0433 0.0450 0.0441 4 0.4898 0.0451 0.0414 0.0320 5 0.5750 0.0687 0.0697 0.0298 6 0.6419 0.0842 0.0858 0.0286 7 0.7157 0.1110 0.1099 0.0267 8 0.8374 0.2100 0.2085 0.0238 
下载: 导出CSV
表 2 计算精度与流量结构的关系(2)
Table 2. Simulation result of model (5)
比例 RATE (H) ARE (X) ARE (Y) ARE (H) 1∶2 0.7623 0.1127 0.2817 0.0287 1∶4 0.6806 0.0614 0.3038 0.0305 1∶8 0.5573 0.0475 0.3772 0.0390 1∶16 0.3681 0.0283 0.4579 0.0557 2∶1 0.7580 0.2287 0.1098 0.0246 4∶1 0.6695 0.1738 0.0491 0.024 1 8∶1 0.5197 0.2500 0.0292 0.0250 16∶1 0.3828 0.2049 0.0141 0.0247
下载: 导出CSV
表 3 数据有误差时计算精度与流量结构的关系
Table 3. Simulation result of model (5) with sample errors
序号 RATE (H) 无误差 有误差 ARE (X) ARE (Y) ARE (H) ARE (X) ARE (Y) ARE (H) 1 0.1675 0.0426 0.0444 0.1638 0.0386 0.0407 0.1625 2 0.2118 0.0495 0.0488 0.1600 0.0400 0.0413 0.1554 3 0.3206 0.0684 0.0629 0.1070 0.0545 0.0431 0.1014 4 0.4232 0.1304 0.1131 0.1029 0.0971 0.0770 0.0923 5 0.5581 0.1679 0.1600 0.0834 0.1237 0.0893 0.0726 6 0.6762 0.2885 0.2645 0.0876 0.1936 0.1430 0.0734 7 0.7458 0.4315 0.3998 0.0805 0.2748 0.2240 0.0651 8 0.8109 0.5719 0.5105 0.0840 0.3954 0.3207 0.0705
下载: 导出CSV
-
[1] Bell M G H. The estimation of origin-destination matrices by constrained generalized least squares[J]. Transportation Research, 1991, 25B(1): 13—22. [2] Bell M G H. The real-time estimation of origin-destination flowin the identification of origin-destination flows[J]. Transportation Research, 1991, 25B(2): 115—125. [3] SHI Feng, LI Mao-sheng. Recursive estimation for origin-destination matrix though a region[A]. Proceedings of the Third International Conference on Traffic and Transportation Studies(ICTTS'2002)[C]. Guilin, China. American Society of Civil Engineers(ASCE), 2002. [4] LI Baibing, Moor B D. Recursive estimation based on the equality-constrained optimization for intersection origin-destination matrices[J]. Transportation Research, 1999, 33B(3): 203—214. [5] Nihan N, Davis G. Recursive estimation of origin-destination matrices from input/output counts[J]. Transportation Research, 1987, 21B(2): 149—163. [6] Sherali H D, Arora N, Hobeika A G. Parameter optimization methods for estimating dynamic origin-destination trip-tables[J]. Transportation Research, 1997, 31B(2): 141—157. [7] Fuller W A. Measurement Error Models[M]. New York: wiley, 1987. [8] 林飞, 曾五一. 对自变量为随机变量的回归模型估计方法的探讨[J]. 统计研究, 1999, 32 (12): 22—27. https://www.cnki.com.cn/Article/CJFDTOTAL-TJYJ199912004.htmLIN Fei, ZENG Wu-yi. Some new way to estimate the parameters of the linear model[J]. Statistical Research, 1999, 32(12), 22—27. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-TJYJ199912004.htm -
点击查看大图
图(2) / 表(3)
计量
- 文章访问数: 225
- HTML全文浏览量: 63
- PDF下载量: 234
- 被引次数: 0
