Advancing Dynamic Relief Response: Integration of New Data Streams and Routing Models

The response to the January 2010 Haitian earthquake demonstrated that a new era of disas- ter relief characterized by an unprecedented use of digital technology to collect and disseminate information had begun. The years since the earthquake have seen a deepening understanding of the roles new technologies are already playing in disaster relief, as well as the severe challenges involved in adequately putting these technologies to use. The goal of this project is to assess how this rapidly growing availability of new types of data shortly after a disaster or during a crisis can lead to better logistical decisions, and to facilitate integration of the data into humanitarian logistics operations, such as relief distribution and search-and-rescue e�orts. The proposed work follows and greatly expands on a research design implemented successfully in our preliminary study on the use of text message data for urban search and rescue (USAR) teams in Haiti. Over the past year we have conducted a preliminary study centered on incorporating real- time information on road conditions, collapsed infrastructure and assistance requests from trapped people into search-and-rescue operational models. The ultimate goal of such work is to provide humanitarian assistance organizations with systematized, deployable methods of putting new data streams directly to use on the ground while working within these data streams' limitations. In the proposed work for the next three years, we plan to scale up and extend our work in this �eld by introducing a new class of models, called adaptive orienteering problems with stochastic travel times, that combine vehicle routing and optimal path �nding in a dynamic framework subject to uncertainty.