交通流问题文献综述(2)

关于短时交通流预测和ITS关系：

Cheslow et al. (1992) underlined the relation of dynamic traffic control to the ability to make and continuously update traffic predictions.

• Cheslow M , Hatcher S G , Patel V M . AN INITIAL EVALUATION OF ALTERNATIVE INTELLIGENT VEHICLE HIGHWAY SYSTEMS ARCHITECTURES[J]. System Architecture, 1992.

关于时间粒度：

Ishak and Al-Deek (2002) concluded that the prediction accuracy degrades as the forecasting horizon increases.

• Ishak, S. and Al-Deek, H. (2002) Performance of short-term time series traffic prediction model, Journal of Transportation Engineering, 128(6), pp. 490–498.

关于方法：

历史平均法：

• Smith, B. L. and Demetsky, M. J. (1997) Traffic flow forecasting: comparison of modelling approaches, Journal of Transportation Engineering, 123(4), pp. 261–266.

平滑方法：

Smith and Demetsky, 1997; Williams et al., 1998

• Smith, B. L. and Demetsky, M. J. (1997) Traffic flow forecasting: comparison of modelling approaches, Journal of Transportation Engineering, 123(4), pp. 261–266.
• Williams, B. M., Durvasula, P. K. and Brown, D. E. (1998) Urban traffic flow prediction: application of seasonal autoregressive integrated moving average and exponential smoothing models, Transportation Research Record, 1644, pp. 132–144.

卡尔曼滤波：

The potential of this was first demonstrated by Okutani and Stephanedes (1984) who used Kalman filtering in urban traffic volume prediction and then developed an extended Kalman filter to predict traffic diversion in freeway entrance ramp areas.

• Okutani, I. and Stephanedes, Y. J. (1984) Dynamic prediction of traffic volume through Kalman filtering theory, Transportation Research, Part B, 18(1), pp. 1–11

ARIMA家族：

In the early 1990s, autoregressive linear processes such as the auto-regressive integrated moving average (ARIMA) family of models, which were first introduced in traffic forecasting by Ahmed and Cook (1979) and Levin and Tsao (1980), provided an alternative approach based on the stochastic nature of traffic. Davis et al. (1991) applied a single auto-regressive integrated moving average (ARIMA) model to forecast the bottleneck formulation in a freeway. Later, Hamed et al. (1995) applied an ARIMA model to forecast urban traffic volume.

• Levin, M. and Tsao, Y.-D. (1980) On forecasting freeway occupancies and volumes, Transportation Research Record, 773, pp. 47–49.
• Davis, G. A., Niham, N. L., Hamed, M. M. and Jacobson, L. N. (1991) Adaptive forecasting of freeway traffic congestion. Transportation Research Record, 1287, pp. 29–33.

非参数方法优于参数方法：

Non-parametric regression is based on the principles of pattern recognition and chaotic systems (Smith et al., 2002). Smith et al. suggested that although traffic conditions are considered as stochastic processes, the chaotic behaviour of traffic conditions near congestion can be modelled by non-parametric regression. The main purpose is to identify clusters of data with behaviour similar to current traffic state at a certain forecasting interval. Non-parametric regression possesses a number of advantages such as its intuitive formulation, the lack of a need for assumption on the transition of traffic states from one period to another, and, finally, the simplicity in modelling multivariate settings (Clark, 2003).

• Smith, B. L., Williams, B. M. and Oswald, K. R. (2002) Comparison of parametric and non-parametric models for traffic flow forecasting, Transportation Research Part C: Emerging Technologies, 10(4), pp. 303–321.
• Clark, S. (2003) Traffic prediction using multivariate nonparametric regression. Journal of Transportation Engineering, 129(2), pp. 161–168.

KNN方法：

In general, the literature shows promising results when using non-parametric regression. First, Smith and Demetsky (1996) tested the performance of nearest neighbour non-parametric regression compared with neural networks, a historical average and the ARIMA model and concluded that the first was superior in the field of transferability and robustness compared with different data sets. Smith et al. (2000) used kernel neighbourhoods and suggested that the method produced predictions with an accuracy comparable with that of the seasonal version of an ARIMA model. Finally, Smith et al. (2002) tested the performance of non-parametric regression based on heuristically improved forecast generation methods and found that this approach did not produce better predictions than seasonal ARIMA. Nevertheless, they supported the fact that a combined model could be used in cases where the requirements of the seasonal ARIMA could not be met. More recently, Clark (2003) found that non-parametric regression was more accurate when predicting flow and occupancy in motorways than speed.

• Smith, B. L., Williams, B. M. and Oswald, K. R. (2000) Parametric and nonparametric traffic volume forecasting, in: Proceedings of the 80th TRB Annual Meeting, Washington, DC.
• Cheng S , Lu F , Peng P , et al. Short-term traffic forecasting: An adaptive ST-KNN model that considers spatial heterogeneity[J]. Computers Environment & Urban Systems, 2018.
• Guo F , Polak J W , Krishnan R . Predictor fusion for short-term traffic forecasting[J]. Transportation Research Part C: Emerging Technologies, 2018, 92:90-100.

SVM方法

• Sun Z , Fox G . Traffic Flow Forecasting Based on Combination of Multidimensional Scaling and SVM[J]. International Journal of Intelligent Transportation Systems Research, 2014, 12(1):20-25.

RF方法

• Hamner B . Predicting Travel Times with Context-Dependent Random Forests by Modeling Local and Aggregate Traffic Flow[C]// IEEE International Conference on Data Mining Workshops. IEEE Computer Society, 2010.