Calibration and Validation of a Regional-Level Traffic Model for Hurricane Evacuation

One of the key issues with evacuation models, or any traffic model, is that some public officials place little faith in the results generated by these models. The way to reassure doubts is by thorough calibration and validation of the model. However, there has been little research in this particular area of evacuation modeling. In an effort to expand the literature on evacuation model calibration/validation, the purpose of this paper is to provide a narrative in which the authors discuss the calibration and validation of a regional-scale evacuation model using observed evacuation data. A traffic simulation model was developed to analyze traffic conditions associated with a regional-level mass emergency evacuation for Southeastern Louisiana. The model was constructed using the TRANSIMS transportation modeling system. The narrative details three important aspects of traffic modeling: trip assignment convergence, macro-level calibration, and network flow validation. Similar results were found after testing two convergence criteria sets. A “descending” criteria set was ultimately chosen since it could lead to a more robust routing solution. During the calibration procedure, it was difficult to determine which parameter value would be the most appropriate. This was an expected result since the calibration focused on the entire regional network rather than select routes. The validation procedure revealed the importance of including background traffic in the simulation model. The results here should lead to further enhancements made toward effectively calibrating and validating traffic models for evacuation analysis.