摘要：Based on the decomposition of the evolution processes of the urban expressway capacity after traffic accidents and the influence factors analysis, an approach for estimating the capacity has been proposed. Firstly, the approach introduces the Decision Tree ID algorithm, solves the accident delay time of different accident types by the Information Gain Value, and determines congestion dissipation time by the Traffic Flow Wave Theory. Secondly, taking the accident delay time as the observation cycle, the maximum number of the vehicles through the accident road per unit time was considered as its capacity. Finally, the attenuation simulation of the capacity for different accident types was calculated by the VISSIM software. The simulation results suggest that capacity attenuation of vehicle anchor is minimal and the rate is 30.074%; the next is vehicles fire, rear-end, and roll-over, and the rate is 38.389%, 40.204%, and 43.130%, respectively; the capacity attenuation of vehicle collision is the largest, and the rate is 50.037%. Moreover, the further research shows that the accident delay time is proportional to congestion dissipation time, time difference, and the ratio between them, but it is an inverse relationship with the residual capacity of urban expressway.
