Citation

Hou L, Lao Y, Wang Y, Zhang Z, Zhang Y, Li Z. Transp. Res. C Emerg. Technol. 2013; 28: 87-100.

Copyright

(Copyright © 2013, Elsevier Publishing)

DOI

10.1016/j.trc.2012.12.005

PMID

unavailable

Abstract

Incident response time is critical for incident management. The sooner an incident is responded to, the lower the negative impact comes from it. There have been some achievements on incident response time modeling. However, most of them were based on empirical observations rather than the mechanism of the system and hence their findings were highly dependent on the proposed hypotheses and study sites. A more general analytical method is needed for response time analysis. To fill up the gap, a mechanism based approach is proposed to model the incident response process and explore the contributing explanatory attributes in this paper. A typical incident response process is mathematically formulated based on the incident response truck (IRT)'s activity. Response time is considered being comprised of both preparation delay and travel time to the incident site. Both components are modeled using probability distributions to take their stochastic features into account. The response time model is calibrated using the Washington State Incident Tracking System (WITS) data and dual-loop detector data collected in 2009. Seven variables were found to significantly increase the response preparation delay (e.g. injury involved, heavy truck involved, and weekends) and eleven variables were found having a decreasing effect on the preparation delay (e.g. peak hour and average annual daily traffic). The model has the potential to be used for incident response resource optimization and identification of measures for incident response time improvement.