Legion is a a suite of AI agents trained on numerous simulated battles against an optimized enemy and Enemy COAs (ECOAs) to provide operational payoff in all measures of performance including reduced inventory usage and reduced cost per kill.

The Red Force COA generator (the adversary) AI agents learn and optimize to defeat Blue defenses, forcing further improvement in the Blue DP as it adapts to an increasingly smart enemy. DRL/AL algorithms learn and generate COAs to maximize utility of all available sensors and weapons to optimize success against threats. As the Blue DP learns, it exploits knowledge gained through exploration of the problem environment to inform and improve the long-term payoff of future decisions. The Blue DP converges to an optimal decision policy by learning from Blue behavior an optimized Red Force strategy.

Current Protocols For Battle Management Fall Short.

  • Unresponsive battle management tools capable of understanding the threat scene for complex multi-weapon and multi-threat type engagements.
  • Inefficient tasking of sensors to optimize performance.
  • Inability to manage timelines and kill chain needs for system level engagement.
  • Business logic-based plans fall short leading to decreased battlespace, resource waste, and leaked threats with the resulting damage and asset loss.
  • Traditional dynamic planners rely on optimization schemes and user-tuned objective functions such as integrated QoS, time to fire control quality track, time to “discrim complete”, fraction of lethality negated, Single Shot Engagement Kill Probability (SSEKP), PES, and others.
  • Timely decision making and system effectiveness suffers.

Legion™ consists of three components.

Reinforcement Learning

Builds optimal decision policy based on actions and rewards. (training side)

Specialized Expert Agents

Trained for optimal performance on specific weapon and sensor elements that make decisions for their element in support optimal system performance.


Scenario builder allows the user to view areas, threats, and scenarios, to execute scenarios, and playback to analyze the scenarios.

Request A Demo of Arcarithm's Command & Control Solutions.

Arcarithm is in the business of developing innovative solutions to solve the hardest problems. Please get in touch to learn how we can add value to your mission.

1002 Meridian Street N
Huntsville, Alabama 35801


Customer Service

Got a question?

Need a technical briefing?

Learn how Arcarithm can deliver high rewards at lower risk, and add value to your mission.


Or Send Us A Message