Dynamic Plan Evaluation for Military Logistics

The defense personnel engaged in logistics planning need tools that allow them to evaluate a plan and efficiently update this information as the situation evolves. This paper reports on the work that IET performed as part of DARPA's Ultra*Log project to design a probabilistic reasoning algorithm for dynamic plan evaluation that conforms to the stringent performance requirements of Ultra*Log's multiagent planning architecture called Cougaar. Introduction & Background The ability to dynamically evaluate the effects of changes in the world situation on the quality of a logistics plan is crucial for military commanders who have to decide if a plan is feasible in a given situation. Furthermore, when multiple courses of action are feasible, they have to decide on the best choice for execution in a situation. The decision makers involved in military logistics planning need tools that can quickly evaluate a logistics plan given uncertain and incomplete information about a situation. Previously, we have reported on the facility that we developed to generate and manage multiple courses of action in Cougaar−a hierarchical planning, execution monitoring, and replanning system developed to solve military logistics planning problems (Upal). Cougaar (Cognitive Agent Architecture) (BBN 2002) is a multi- agent system developed under DARPA's Advanced Logistics Project (ALP) and its successor Ultra*Log. This paper reports on the work that IET performed as part of Ultra*Log to add the ability to efficiently evaluate the effects of changes on the quality and feasibility of a course of action and choose the best course of action in an uncertain situation when multiple plans are available. Cougaar society closely mirrors the organizational structure of the military units engaged in real-world