Component-Based Simulation and Agent-Based Brokering: Towards Ad Hoc Simulations in Crisis and Emergency Management

Crisis and emergency management systems often require the use of large simulations to predict future developments of the simulated phenomena. For example, floods and forest fires are dynamic phenomena with devastating consequences. Simulations for these disasters require large amounts of information, including current and future weather data and the topography of the local terrain. In both cases, the ability to create an efficient simulation to support responses to a crisis is useful only if the development time is short enough and the simulation is fast enough to be used in real-time crisis management. This paper introduces our work in developing software tools for integrating simulations and sensor data in support of crisis and emergency management. In particular, it presents three basic challenges for developing these tools and discusses our approaches for meeting these challenges.