Decision Superiority

Executive Summary

We will develop an artificial-intelligence-powered decision support technology for strategic wargaming through high-performance-computing-enabled conflict simulation, intelligence-informed adversary modeling, and optimal control algorithms. Through careful experimentation utilizing realistic conflict simulations, we will develop an important new computational capability for generating creative solutions to multi-domain deterrence scenarios.

Publications, Presentations, and Patents

ABMARL. Agent-based modeling tool for python developed by Ephraim Rusu. https://github.com/LLNL/Abmarl

PPOAF. MPI-enabled reinforcement learning algorithms by Alister Maguire. https://github.com/aowen87/ppo_and_friends

GISKARD. Environmental data aggregator and light-weight simulation engine by Alexx Perloff.

Orbit-gym. Orbital reinforcement learning environments built on top of LLNL’s ssapy by Alister Maguire.

JCATS+. Python interface/client for JCATS conflict simulation software by Ephraim Rusu, Abel Gezahegne, Nick Matyas, and Joe Hennis.

Dan Merl and Michale Schneider, (Presentation, Decision Superiority Project Briefing, US STRATCOM,30 November 2021).

Dan Merl, "Towards Cognitive Deterrence" (Presentation, Protocol Visit Briefing at LLNL to Director of DOE AITO, Livermore, CA,  2 March 2022).

Merl, Dan. Towards Cognitive Deterrence. LLNL Cognitive Simulation Workshop. 15 March 2022. LLNL-PRES-833726.

Dan Merl and Michael Schneider, "Reinforcement Learning for Understanding Tactics and Strategy" (Presentation, Protocol Visit Briefing at LLNL to visitors from US State Dept., Livermore, CA, 23 May 2022).

Dan Merl and Michael Schneider, "Reinforcement Learning for Understanding Tactics and Strategy" (Presentation - Protocol Visit Briefing at LLNL to additional visitors for US State Dept., Livermore, CA, 15 August 2022).

Dan Merl and Michael Schneider, "Decision Superiority Overview"(Presentation, Strategic Capabilities Office Briefing, Livermore, CA, 4 November 2022).

Dan Merl, "Toward Machine Assisted Decision Superiority" (Presentation, DOD Advisory Board,  Livermore, CA, 9 Feb 2023).

Dan Merl, "Decision Superiority" (Presentation, Computing External Review Committee, Livermore, CA, 18 April 2023).

Dan Merl, "Next generation decision support through AI-enabled conflict assessment" (Presentation - Congressional protocol visit, Livermore, CA, 24 May 2023).

Ephraim Rusu, "AI-enabled conflict simulation" (Presentation, Sandia Machine Learning and Deep Learning Workshop, Livermore, CA, 20 July 2023).

Michael Schneider, "Decision Support" (Presentation, NNSA Strategic Futures Summit, Sandia National Laboratory, August 9 2023).

Dan Merl and Michael Schneider, "Decision Superiority Initiatives at LLNL" (Presentation - Intelligence Community visitors, Livermore, CA, 24 August 2023).

Michael Schneider, "Decision Superiority" (Presentation, USSTRATCOM Science & Technology Forum, September 14, 2023).