A Dynamic Constraint Reasoning Approach to Mixed Initiative Inter-Modal Lift Planning Theory

Missions for Special Operations Forces (SOF) are distinctive in their heavy joint service participation, their strict timing constraints, and their peculiar equipment requirements. Lift support planning for multiple SOF missions requires flexible assetscheduling, intelligent asset tracking and routing, and multiple-resource problem-solving. Fully automated planning, though, is inefficient and inappropriate for exploring all the options for meeting lift commitments.In Phase I, we explored a knowledge-based approach for modeling the unique aspects of the SOF IML planning domain, representing SOF lift plans as interval-based dynamic constraint networks and developing a novel mixed-initiative planning process togenerate consistent and complete lift plans.In Phase II, ALPHATECH proposes to implement a mixed-initiative planner that facilitates plan synthesis, evaluation, and repair. This planner will intelligently guide the user through plan synthesis by filtering out inconsistent planning options. Bytransforming partial plans into a dynamic constraint network, based on our constraint-based interval representation of the problem domain, the software can reason about temporal and metric resource constraints during continually evolving plan development.We will use path-consistency algorithms on the constraint network to detect flawed partial plans, and heuristically propose possible repairs and alternate plans.