You are here

Warfighting Chess Games and Pieces

Description:

TECHNOLOGY AREA(S): Info Systems, Human Systems 

OBJECTIVE: Objective is to mature a simulation capability that can play smart red forces against smart blue forces in order to develop decision support tools and reverse engineer legacy simulator entity behavioral and models to support “fair fights” (1). Entity behavior and models and one or more decision support tools will be “built” and verified by allowing an artificial intelligence (AI) capability to watch thousands of “games” played between smart agents (1). A condition when the differences between the performance characteristics of two or more interoperating simulations have significantly less effect on the outcome of a simulated situation than the actions taken by or resources available to the simulation participants. 

DESCRIPTION: The goal of the topic is to develop warfighting decision support tools and reverse-engineering processes by allowing AI technology to observe automated smart red forces compete with smart blue forces within a simulation. Just as computer-based chess games are able to make optimal moves given the current state of the board and the most likely future state, a military decision support aid should suggest modifications to current plans and predict future outcomes given current content of the common tactical and intelligence picture. It is expected but not required that deep learning be used to learn optimal actions relative to a set of measures of effectiveness (MOEs) and measures of performance (MOPs). Reverse-engineering processes are needed to develop accurate behavioral and entity models in legacy simulations. For example, over the past decade the Marine Corps has procured a number of individual simulator systems that are either proprietary in nature, or have dissimilar behavioral and entity models. As the Marine Corps moves towards a common architecture within a live, virtual, and constructive paradigm it is necessary that entities behaviors and models are normalized to ensure fair fights. Unlike chess boards, there are different representations of environments within each simulation that are used during virtual and constructive simulation. In is critical for training and wargaming that these representations have similar entity behaviors and models. To generate adequate and relevant training data, the behaviors and models of actors in the simulation need to be controlled by smart agents that can run much faster than real time. The Phase I effort will be somewhat bounded, involving three agent platoons against three opposing platoons, each trying to secure an area of interest. Initial positions for the six units should be random. Other random variables should include weather, maneuver obstacles, terrain, equipment failures, and readiness levels. Both virtual and constructive simulators must be used. Performers can be sponsored for simulations (e.g., OneSAF) mission simulator if desired and needed. Mission simulators are increasingly capable in terms of programmable behaviors and interactions. While AI software has been demonstrated to beat humans at games of increasing complexity (Tic-Tac-Toe to Go), agents have not been developed that can support decision making in the context of a military mission. Confounding this challenge, there are several legacy Marine Corps simulators that exist with different entity behaviors and models that are not easily configurable to interoperate. Specific technical challenges include model normalization and model development relevant to a high-dimensional, continuous state space involving an adversarial agent. Work produced in Phase II may become classified. Note: The prospective contractor(s) must be U.S. Owned and Operated with no Foreign Influence as defined by DOD 5220.22-M, National Industrial Security Program Operating Manual, unless acceptable mitigating procedures can and have been implemented and approved by the Defense Security Service (DSS). The selected contractor and/or subcontractor must be able to acquire and maintain a secret level facility and Personnel Security Clearances, in order to perform on advanced phases of this contract as set forth by DSS and ONR in order to gain access to classified information pertaining to the national defense of the United States and its allies; this will be an inherent requirement. The selected company will be required to safeguard classified material IAW DoD 5220.22-M during the advance phases of this contract. 

PHASE I: Determine feasibility for the development of an operationally relevant model normalization tools and a model based decision support tool. Conduct a detailed analysis of literature and commercial capabilities. For a bounded number of legacy models, actors, and behaviors, conduct a lab-based proof of concept demonstration. During the Phase I effort, performers are expected to identify metrics to verify performance of model normalization and decision support tools with the goal of reducing technical risk associated with building a working prototype, should work progress. Performers should produce Phase II plans with a technology roadmap and milestones for prototype development. 

PHASE II: Produce a prototype system based on the preliminary design from Phase I. The prototype should enable human users to compete against agents or agents against each other within relevant Naval mission simulations. The system must be able to bring in legacy models for specific behaviors and entities. Additionally, the system must provide explanatory evidence for decision recommendations in terms of extrapolated measures of performance/effectiveness. The performance of agents will be measured by comparing simulation outcomes. During Phase II, the small business may be given specific scenarios by the Government to validate capabilities. An offeror should assume that the prototype system will need to run as a distributed application with a mature design for the human computer interface. Phase II deliverables will include a working prototype of the system (source code and executable), software documentation including a user’s manual, and a demonstration using a Naval operational scenario of interest. It is probable that the work under this effort will be classified under Phase II (see Description section for details). 

PHASE III: Produce a final prototype capable of deployment to training centers, operational command and control centers, and as a virtual application. The system should be adapted to transition as a component to a larger system or as standalone commercial product. The small business, working with transition and commercialization partners, should provide a means for performance evaluation with metrics for analysis (e.g., accuracy of decision support) and method for operator assessment of product interactions (e.g., display visualizations). The Phase III system should have an intuitive human computer interface. The software and hardware should be modified and documented in accordance with guidelines provided by engaged programs of record and commercial partners. Researchers are encouraged to publish Science and Technology (S&T) contributions. Technology development will be applicable to the commercial gaming market, particularly to games with adversarial interaction involving a large, continuous feature space. The private sector also faces similar challenges with hierarchical modeling architectures that include legacy products. 

REFERENCES: 

1: Abar, S., et al. "Agent Based Modelling and Simulation tools: A review of the state-of-art software", Computer Science Review 24 (2017) 13–33

2:  2. Liebowitz, J. "Sharing the Solution", Computers ind. Engng 16, NO. 4, pp. 587-593, 1989

3:  Pan, Y. "Heading toward Artificial Intelligence 2.0", Engineering 2 (2016) 409–413

4:  Brynielsson, J. "Using AI and games for decision support in command and control", Decision Support Systems 43 (2007) 1454–1463

5:  DoD Modeling and Simulation Glossary. accessed on 25 June 2017. https://www.msco.mil/MSReferences/Glossary/TermsDefinitionsE-H.aspx

KEYWORDS: Artificial Intelligence; Warfighting Simulation; Modeling; Training; Agent Based Models; Gaming; Decision Support 

CONTACT(S): 

Martin Kruger 

(703) 696-5349 

martin.kruger1@navy.mil 

Peter Squire 

(703) 696-0407 

Frank Segaria 

(202) 404-8988 

US Flag An Official Website of the United States Government