Objective: Develop automated algorithms to help MCMCs successfully respond to changing operational conditions and performance anomalies in minutes as opposed to hours.
Our Approach: Knexus scientists and engineers developed computationally efficient AI algorithms for automating situation monitoring, scheduling and rescheduling to provide valid and focused tactical decision support to MCM commanders and staff. They utilized their prior experience and pioneering research on AI planning algorithms to extend and apply a variety of hierarchical automated planning techniques. These algorithms generate diverse schedules options in minutes. They also developed a new algorithm for “minimally disruptive” plan repair that encoded lessons learned from field experience using a memory-based machine learning technique to quickly respond with reschedule options.
They demonstrated and evaluated these algorithms with an advanced system prototype. This prototype, implemented in a service oriented architecture, provided tactical planning, monitoring, and replanning support and visualization in a browser. Field experts validated the approach as sound and valid by performing walkthroughs and simulated decision making.
Solution We Delivered: The proof-of-concept prototype and the included AI algorithms and technologies were demonstrated at Technology Readiness Level (TRL) of 5. This substantially reduced the technical risk and gave the sponsors the necessary confidence and basis to proceed with a concept of operations for a next generation MCM operational tactical decision support tool. The findings and the success of this project has resulted in a follow-on award for Knexus under a Future Naval Capabilities Program. In this program, Knexus will further develop and mature the technologies into a robust, and user friendly, decision support system. The following technical and popular press articles were published.
Key Concepts: Tactical Decision Support, AI Enabled Automated Planning, MCM Operational Planning,