中文摘要：

依托从英国快速路系统中采集到的实际交通流数据，研究了时间汇集间隔对交通流异方差性的影响．使用1-30min共30种时间汇集间隔，生成了30个实际交通流数据序列，确定并估计了相应的ARIMA模型，计算后得到30个交通流量残差序列．针对不同汇集间隔下的ARIMA模型残差序列，应用portmanteau Q检验和LM（Lagrange multiplier）检验，分析了交通流量序列的异方差性．实证结果表明：交通流量序列在选定的30个汇集间隔都具有显著的异方差性；较长的时间汇集间隔可以消减交通流量序列中的噪声，从而减弱交通流异方差性的程度．研究结果有助于开发具有较高可靠性和鲁棒性的交通管理和控制系统．

英文摘要：

The effect of the aggregation interval on vehicular traffic flow heteroscedasticity is investigated using real-world traffic flow data collected from the motorway system in the United Kingdom. 30 traffic flow series are generated using 30 aggregation intervals ranging from 1 to 30 min at 1 min increment, and autoregressive integrated moving average （AR/MA） models are constructed and applied in these series, generating 30 residual series. Through applying the portmanteau Q-test and the Lagrange multiplier （LM） test in the residual series from the ARIMA models, the heteroscedasticity in traffic flow series is investigated. Empirical results show that traffic flow is heteroscedastJc across these selected aggregation intervals, and longer aggregation intervals tend to cancel out the noise in the traffic flow data and hence reduce the heteroscedasticity in traffic flow series. The above findings can be utilized in the development of reliable and robust traffic management and control systems.