Abstract Information Infrastructures for Crisis Management
This project carries out research on fundamental computer science problems that arise in the development of information infrastructures for preventing or managing crises such as terrorist attacks, pandemics, chemical spills, hurricanes and earthquakes. Crises have elements of the unexpected despite the planning that goes into preparing for them. The element of surprise implies that crisis-management platforms are on-demand systems that are created when the crisis strikes and that are modified as requirements change. The on-demand aspect of crisis management implies that task force members, other than information technology specialists, must be able to specify and deploy software agents that help the task force manage the crisis. Since the membership of a task force dealing with a crisis is often not known until the crisis strikes, the underlying computing and communication systems available to the task force may not be known until the task force is formed.
This project develops platforms that can be used to implement crisis-management infrastructures on top of heterogeneous systems used by different institutions in a task force. The infrastructure must be able to deal gracefully with changing coalition membership and changing amounts and types of computing resources by adjusting the net processing rate and the machines on which agents are deployed. The project develops theories, methods and tools to help deal with this uncertainty by making agents and messages persistent, stored in databases or file systems in persistent storage systems, and theories and prototypes for dynamic creation of agents and continuous evolution of compositional structures. The project integrates research in many fields to develop infrastructures that support crisis management.
