Travel time reliability for recovery activity immediately after disaster

AbstractThere are many models on travel time reliability that have been applied to many traffic conditions including traffic jams and road constructions (Wakabayashi, 2008). In fact, travel time reliability is also important for disaster recovery (Aaskura et al., 1998). This paper proposes a concept and a requirement for a new travel time reliability index in disaster recovery period.The traditional connectivity reliability analysis can not reflect the travel time reliability in disaster areas although the travel time is important in disaster recovery period. Thus a new index which reflects the travel time reliability should be developed in order to insure the emergency operation (Wakabayashi, 2008). We develop a new model for travel time reliability to relieve the excess demand of traffic when a disaster occurs.The travel time reliability for disaster recovery in our model is different from the usual travel time reliability because of the special factors in disaster. Firstly, the traffic capacity is not considered in our model because the road network would be controlled by transportation department and the traffic capacity can be ignored under the human control. Secondly, the buffer time in our model is difficult to estimate because the traffic condition is unknown when a disaster occurs (Lomax et al., 2003), so the traffic accessibility is considered in our model (Lwan et al., 2003). Thirdly, the path's distance is an important factor in our model. The influence of path's distance will be remarkable in the central disaster areas and almost disappear in the boundary part of disaster areas. It also means that the influence of distance will increase no longer when the length of the distance is beyond a certain critical point.The relation between travel time reliability for disaster recovery and traffic accessibility is also studied from various network hierarchies, involving path, OD pair and the whole network (Tu et al., 2007). The parameters which stood for the disaster scale or the damage level are also discussed.Because the disaster of some areas is unpredictable, we only use the historical data to certificate the model, and then the simulated results of our model for some disasters are imperfect. So this model is only a theoretical model and cannot be used for the actual situation. More factors and simulations should be considered to solve the defects of our model