Stilman B.,STILMAN Advanced Strategies |
Stilman B.,University of Colorado at Denver
International Journal of Machine Learning and Cybernetics | Year: 2014
The hierarchy of formal languages is a mathematical representation of linguistic geometry (LG). LG is a type of game theory for a class of extensive discrete games called abstract board games (ABG), scalable to the level of real life defense systems. LG is a formal model of human reasoning about armed conflict, a mental reality "hard-wired" in the human brain. LG, an evolutionary product of millions of years of human warfare, must be a component of the primary language of the human brain (as introduced by Von Neumann). Experiences of development of LG must be instructive for solving another major puzzle, discovering the algorithm of discovery, yet another ancient component of the primary language. This paper reports results on discovering mental processes involved in the development of the hierarchy of formal languages. Those mental processes manifesting execution of the algorithm of discovery are called visual streams. This paper reveals the visual streams that were involved in the thought experiments led to the development of the formal theory of LG. Specifically, it demonstrates the streams involved in choosing the formal-linguistic representation of LG; the type of formal languages and grammars, the so-called controlled grammars; the construction of the grammars of shortest trajectories and the grammar of zones. This paper introduces a hypothesis of how we construct and focus visual streams. © 2013 Springer-Verlag Berlin Heidelberg. Source
Agency: Department of Defense | Branch: Navy | Program: SBIR | Phase: Phase I | Award Amount: 70.00K | Year: 2003
We propose to develop and demonstrate a new architecture for engineering human understanding into automated combat systems. This architecture will provide a conceptual framework for human-centered automated combat systems satisfying reduced-manningrequirements of the transformed future Navy. The approach is based on Linguistic Geometry (LG), a mathematical theory for strategic superiority discovery. The most significant advantages of the LG approach are modeling of the intelligent enemy andextraordinarily fast automatic generation of best strategies, tactics and COA for all the sides of a conflict. We will also develop the operational specifications of an LG-based software tool, LG-SEAGUARD, representing management of automated combatsystems processing within a reduced-manning CIC and aviation command and control nodes using the above architecture. LG-SEAGUARD will reflect the specifics of the Navy and Joint operations at various levels of resolution. It will generate the beststrategies and tactics for all sides of a conflict by optimizing them against various criteria to reflect different types of military operations. LG-SEAGUARD will be dynamic and adaptable to the strategy and tactics of its human adversary. It will plan theoperation, allocate resources with minimal cost to achieve certain probability of success (defined by a planner), generate and assess possible courses of actions (COA) for all sides, select counteractions, control operation by re-planning in real time, andplay and re-play various
Agency: Department of Defense | Branch: Air Force | Program: SBIR | Phase: Phase I | Award Amount: 100.00K | Year: 2005
We propose to develop LG-WARGAMER, a new an asymmetric adversary simulation environment that can support existing military simulations such as the Joint Synthetic Battlespace (JSB) and the Air Force M&S Foundations initiative for Air Force acquisition and training. In Phase I, we will establish feasibility and develop principles for a high level LG-based architecture for such tool. In Phases II and III, a succession of progressively more powerful software prototypes of LG-WARGAMER, implementing the framework developed in Phase I, will be developed. The approach is based on Linguistic Geometry (LG), a new type of game theory changing the paradigms of battle management and mission planning. The most significant advantages of the LG approach are modeling of the intelligent enemy and extraordinarily fast automatic generation of advantageous strategies, tactics, and COA for all the sides of a conflict. LG-WARGAMER shall be capable of interfacing with simulation environments (such as JSB) and providing adversary tactics and strategy to the simulations. Using these capabilities LG-WARGAMER will assist the operators, commanders, warfighters, and trainees in describing, assessing, and predicting the activities of individuals, teams, and organizations. LG-WARGAMER will be capable to consider factors such as the social, cultural, political, economic, religious, ethnic, and ideology of the adversary. Working with JSB, LG-WARGAMER will permit modeling and evaluation of new conceptual military hardware in terms of its functionalities and new strategic and tactical concepts. If a hardware functionality would have hidden flaws (e.g., too low bandwidth of the communication network), the simulated enemy guided by the LG strategies would be able to exploit them providing the hardware evaluators with hands-on proof of failure. Moreover, assisted by the analyst, LG-WARGAMER will discover the direction of changes toward correcting the flaw. Contrariwise, if the functionality has spectacular advantages, LG-WARGAMER would be able to convincingly demonstrate how these advantages could be translated into victory for the Blue forces.
Agency: Department of Defense | Branch: Army | Program: STTR | Phase: Phase I | Award Amount: 100.00K | Year: 2005
We propose to investigate feasibility and develop specs for a new training tool, LG-COACH geared toward distributed interactive training. The tool will be based on Linguistic Geometry (LG). The most significant advantages of LG are a unified conceptual model of all types of military operations, a faithful and scalable model of intelligent enemy, and extraordinarily fast automatic generation of best strategies and tactics for all the sides of a conflict, including the operations with highly divergent and asymmetrical (non-zero-sum) goals of the opponents. In Phase I, we will develop theoretical foundations, operational specifications, and a pilot demo of LG-COACH (to be fully developed on Phase II). LG-COACH will provide the Army with a realistic tool for implementing an effective embedded training system for the development of tactical skills across various levels of command. It will support the TLAC ("Think Like A Commander") approach to training and reusable, adaptable, and scalable vignette-based embedded battlefield decision-making training exercises. It will be able to capture and represent cultural and social idiosyncrasies of conflicts for different parts of the world. It will generate the best strategies and tactics for either side of a conflict (or for all of them) by optimizing them against various criteria to reflect different types of military operations including asymmetric warfare, effects based operations, and future types of warfare. LG-COACH will be truly dynamic, adaptable to the strategy and tactics of its human adversary. Moreover, LG-COACH will provide training for planning the entire operation, allocating resources with minimal cost to achieve certain probability of success (defined by a planner), development of advantageous courses of actions (COA, selecting counteractions, etc. LG-COACH will create the ultimate learning environment for warfighters.
Agency: Department of Defense | Branch: Defense Advanced Research Projects Agency | Program: SBIR | Phase: Phase II | Award Amount: 3.50M | Year: 2007
The goal of this proposal and follow-up development is to apply the Linguistic Geometry (LG) based predictive technology to mission planning and dynamic allocation of operational assets in both Battle Command and Simulation environments. To achieve this goal in Phase II, STILMAN will implement a prototype LG-STRATEGIST enabling the following capabilities with emphasis on integration with FBCB2 and OOS and on evaluation by and transitioning to PEOC3T, PMOOS, RDECOM STTC, and other DoD organizations: a) Rapid construction of new synthetic battlespaces; b) Automatic generation of optimal advantageous strategies for all the sides in a conflict; c) Automatic generation of predictive Red/Blue COA (courses of action); d) Modeling intelligent enemy. The products of this research will be demonstrated in 2 experiments that will assess the contribution of the LG technology to mission planning and the dynamic allocation of operational assets. Ongoing research by the DARPA to increase the utility, (i.e., understanding) of predictive information will be leveraged and focused on the FBCB2 system (Force XXI Battle Command Brigade and Below). By integrating predictive information into the tactical FBCB2 station, comparative experiments will be developed using analysis of variance techniques. As requested by PEOC3T, the experiment process and metrics developed by the DARPA RAID program will be reused and extended to assess the impact of this research on tactical mission performance within upcoming Air Assault Expeditionary Force (AAEF) experiments. While developing LG-STRATEGIST, STILMAN will put a strong emphasis on the development of features reflecting the needs of the sponsoring organizations.