Continuous Interactive Learners for Mission Planning (CILEMP)

A key problem with automated mission planning software is that their underlying domain models may be incomplete, imprecise, inaccurate, or inefficient, and may rapidly become obsolete because of evolving domains. These issues can adversely impact the quality and effectiveness of support offered by automated planning software. In this effort, we will address the planning domain knowledge acquisition problem and capability gap by researching and developing CILEMP, a software system for Continuous Interactive Learners for Mission Planning. Our objective is to develop and deliver a robust and innovative software solution at technology readiness level (TRL) 6 at the conclusion of the Phase II Option period. Our deliverable software system shall enable MCM Commanders to interactively acquire and learn domain models for hierarchical planning in continuous, temporal, uncertain, and adversarial environments. We will accomplish this with four tasks: i) develop CILEMP concept of operations for naval operations focusing on MCM desktop wargaming applications ii) research and develop CILEMP multi-strategy domain model learning algorithms for uncertain, adversarial environments, iv) architect and implement a domain general CILEMP software system that successfully acquires and updates planning domain models, and iv) demonstrate, test, and evaluate CILEMP software for its ability to improve automated operational planning performance.