Operation Design for Incident Response

In order to design an incident response system and the associated dispatch model, we need to understand the role of incident dispatch and the timing to use the decision support system for incident response first. Dispatching a response team is one of the components in freeway-incident duration time. Zografos et al. (1993) shows the total remedy time by Fig. 3.1. The total remedy time is divided into four intervals in this figure. The first interval is T1 “Detection and Identification time”; it is the time interval from the time of incident occurrence to the time that the incident-management center is notified the incident type, position and severity. The sources of notification can be the freeway police, drivers, victims, monitors or even managers of the incident-management center. The second interval is T2 “Dispatch time”; it is the time between incident detection and the time that the incident is assigned to a response team.

The dispatcher in the freeway-management center will firstly use CCTV (a camera on freeway) near the position of the incident to confirm the information from the source after receiving a notification, and then the dispatcher will input the parameters into the dispatch model through a user interface. The model solution will return suggestions regarding response team and route assignment to the dispatcher, and the dispatcher will make a final decision and contact the response team chosen. The third interval is T3

“Preparation and Travel time”; this is the time that the response team needs to arrive at the location of the incident from its previous position. This time is composed of preparation time and travel time. Before the response team departs to the location of the incident, the team should check workers and equipment or complete the duty of the current incident restoration; this period is named preparation time. The last interval is T4 “Clearance time”; this is the time that the response team spends on clearing the

incident at scene. The worker must record the time and the condition at scene and send it back to freeway-management center as the dispatch data via an on-board unit of the incident response system.

Zografos et al. (1993) mentioned that reducing of any component of remedy time can decrease the incident delay on freeways. In this study, the main objective is to minimize the incident response time, which includes T1, T2 and T3, and we focused on how to support the dispatcher to make decisions faster and provide better suggestions that can enable response teams to arrive at the locations of incidents as quick as possible.

Fig. 3.1 Components of Freeway-Incident Duration Time

Based on the background knowledge of freeway-incident duration time, this study designs the operation procedure of the incident dispatch system and shows it in

Fig. 3.2. A scenario that triggers the procedure starts from one or some incident notifications received by the dispatch platform. After the dispatchers check the incident and obtain enough information, they need to input required parameters into the dynamic dispatch model. The source data of the parameters can be divided into three groups, Vehicle Information, Incident Occurrence and Real-time Freeway Network Traffic.

The introduction of these groups of data is in the following:

(1) Vehicle Information

The data of vehicle information contains the status and the position information. The vehicle position information can be collected by GPS devices. The vehicle status information, which indicates if the team is currently on duty, can be acquired through communicating with the response team. In practice, response teams need preparation time to check workers and equipment before they leave their stationary points. On the other hand, if a response team currently processing an incident is assigned another incident, the team needs time to clear the current one. In the model, these two statuses (at the stationary point or processing an incident) of response teams are calculated as waiting time, which means the interval between the time that a response team is assigned an incident and the time that this response team actually departs to the location of incident. The dynamic dispatch model has to calculate the total response time based on the status and position information of response teams upon the moment that the

notification of an incident is received.

(2) Incident Occurrence

Incident information contains severity, positions and estimated processing time of an incident. The severity of incident may be of high, medium and low levels according to the incident type and the conditions at the scene, and the dispatcher will judge the severity level after integrating the information they receive. The position of the incident is related to the travel time in the model, and it is taken as the node in the virtual network used for formulating the problem. This study uses average processing time of each incident type as the estimated processing time in the proposed.

(3) Freeway Network Traffic

The freeway network including interchanges and stationary points of response teams should be constructed first. Traffic data every hour on the freeway are collected and a travel time matrix over the network made by using the shortest path algorithm and traffic data of the network to calculate the travel time between every node pair. The model can retrieve travel times of links through an inquiry function from the matrix when it is building the virtual network of response teams and incidents.

After the dispatcher inputs the value of the parameters, the dynamic dispatch model starts a solution process to determine the optimal assignment. This model seeks to handle all incidents in the freeway network with the minimum response time and must not violate the constraints. This model considers the incident priority order, reassignment and presents them in a minimax problem. The details of notation, formulation and the concept used in the model are explained and discussed in next section.

The output of the model is the dispatch strategy, including the assignment of each incident assigned to an appropriate response team and the associated route. The incident

dispatch result upon the notification of a new incident will be stored and used as the parameters for the problem of the later incident.

Fig. 3.2 Operation Design for Incident Response