Developing spatio-temporal models for retrofit and reconstruction strategy under unscheduled events

This dissertation consists of three papers. The purpose of the first paper is to design conceptually (1) a spatio-temporal data model for analyzing dynamically changing longer term spatial problems, and (2) Spatio-Temporal Analysis Models (STAM) for analyzing network economic loss due to unscheduled events such as earthquake, flood, terrorism, etc. The first part of this paper examines existing spatio-temporal data models and concludes that a feature-based spatio-temporal data model utilizing an object-oriented modeling technique is the most suitable. The second part of the paper proposes the two types of STAMs: (1) STAM for a priori unscheduled event (STAM-1) and (2) STAM for a posteriori unscheduled event (STAM-2) in order to overcome the following common issues: (1) lack of considering the national economic impact, (2) lack of integration between the traffic flow and commodity flow models, and (3) lack of considering the spatio-temporal characteristics of the economy. For these two models to have spatio-temporality, quarterly Sequential Interindustry Model (SIM) is utilized in both models. The second paper is about the implementation of spatio-temporal models for identifying critical network links for retrofit priority. Previous research focused on identification of critical network links for a given year employing static models. However, the results of the previous research do not accurately reflect the spatio-temporal characteristics of the economy. To overcome this issue, this research adopts STAM-1, and proposes an economic significance index to identify critical network links. The third paper is about the implementation of a spatio-temporal model to develop an infrastructure reconstruction strategy under unscheduled events. The reconstruction strategy chosen to repair damaged transportation network infrastructure after an unscheduled event is important to return the disrupted economy to pre-event status. Moreover, the strategy chosen determines how fast the national economy will recover. The optimal sequence or priority for reconstruction of damaged links needs to be developed in order to quickly restore the economy. This paper proposes and implements the framework for finding the optimal reconstruction strategy by using STAM-2 and a genetic algorithm.