Optimization-based Decision Support Algorithms for Network Identification and Dynamic Resource Management

Rapid mission planning, re-planning and execution are major operational issues in highly dynamic, asymmetric, and unpredictable mission environments. Two basic technical challenges in tackling these issues are: 1) transforming data into information: how to distill relevant and accurate information based on imperfect, incomplete and uncertain observations (data); 2) converting information into actionable decisions: how to generate appropriate Courses of Action (COA) based on information on the mission, the environment and the capabilities of resources to achieve the mission objective(s). In this thesis, we tackle these technical challenges by addressing two fundamental problems: 1) the network identification problem of converting message and transaction data into actionable information (e.g., adversarial organizational structure); 2) the resource management problem of allocating heterogeneous assets to interdependent tasks as one of generating actionable decisions based on information. We model both problems mathematically, link them to generalized versions of existing mathematical optimization problems, leverage novel optimization techniques to solve the resulting problems, and embed the resulting solution approaches into a mixed-initiative information and decision support models for collaborative mission planning, execution and monitoring.