"Much work has been done to develop intelligent tutoring systems in domains such as algebra, geometry, and computer programming. Our work is to develop an intelligent tutoring system to train US soldiers. One main difference in this domain is that one of the main skills to be learned is cooperation between teammates, so the tutor must emphasize collaboration as a skill. In addition, to help train this skill the system must be able to run in real-time, and provide both computer generated teammates, as well as intelligent opposing forces. This system is the first real-time, multi-user, model tracing tutor with simulated teammates. The goal of this thesis is to build a prototype system to validate that this is a valid approach for this domain."

In large scale entity-level military force-on-force simulations it is essential to know when one entity can visibly see another entity. This visibility determination plays an important role in the simulation and can affect the outcome of the simulation. When virtual Computer Generated Forces (CGF) are introduced into the simulation these intervisibilities must now be calculated by the virtual entities on the battlefield. But as the simulation size increases so does the complexity of calculating visibility between entities. This thesis presents an algorithm for performing these visibility calculations using Graphical Processing Units (GPU) instead of the Central Processing Units (CPU) that have been traditionally used in CGF simulations. This algorithm can be distributed across multiple GPUs in a cluster and its scalability exceeds that of CGF-based algorithms. The poor correlations of the two visibility algorithms are demonstrated showing that the GPU algorithm provides a necessary condition for a "Fair Fight" when paired with visual simulations.

Context-based Reasoning (CxBR) and Contextual Graphs (CxGs) involve the modeling of human behavior in autonomous and decision-support situations in which optimal human decision-making is of utmost importance. Both formalisms use the notion of contexts to allow the implementation of intelligent agents equipped with a context sensitive knowledge base. However, CxBR uses a set of discrete contexts, implying that models created using CxBR operate within one context at a given time interval. CxGs use a continuous context-based representation for a given problem-solving scenario for decision-support processes. Both formalisms use contexts dynamically by continuously changing between necessary contexts as needed in appropriate instances. This thesis identifies a synergy between these two formalisms by looking into their similarities and differences. It became clear during the research that each paradigm was designed with a very specific family of problems in mind. Thus, CXBR best implements models of autonomous agents in environment, while CxGs is best implemented in a decision support setting that requires the development of decision-making procedures. Cross applications were implemented on each and the results are discussed.

This thesis investigates the effectiveness of employing pretraining and a discrete "control signal" bottleneck layer in a neural network trained in aircraft navigation through deep reinforcement learning. The study defines two distinct tasks to assess the efficacy of this approach. The first task is utilized for pretraining specific parts of the network, while the second task evaluates the potential benefits of this technique. The experimental findings indicate that the network successfully learned three main macro actions during pretraining. flying straight ahead, turning left, and turning right, and achieved high rewards on the task. However, utilizing the pretrained network on the transfer task yielded poor performance, possibly due to the limited effective action space or deficiencies in the training process. The study discusses several potential solutions, such as incorporating multiple pretraining tasks and alterations of the training process as avenues for future research. Overall, this study highlights the challanges and opportunities associated with combining pretraining with a discrete bottleneck layer in the context of simulated aircraft navigation using reinforcement learning. / Denna studie undersöker effektiviteten av att använda förträning och en diskret "styrsignal" som fungerar som flaskhals i ett neuralt nätverk tränat i flygnavigering med hjälp av djup förstärkande inlärning. Studien definierar två olika uppgifter för att bedöma effektiviteten hos denna metod. Den första uppgiften används för att förträna specifika delar at nätverket, medan den andra uppgiften utvärderar de potentiella fördelarna med denna teknik. De experimentella resultaten indikerar att nätverket framgångsrikt lärde sig tre huvudsakliga makrohandlingar under förträningen: att flyga rakt fram, att svänga vänster och att svänga höger, och uppnådde höga belöningar för uppgiften. Men att använda det förtränade nätverket för den uppföljande uppgiften gav dålig prestation, möjligen på grund av det begränsade effektiva handlingsutrymmet eller begränsningar i träningsprocessen. Studien diskuterar flera potentiella lösningar, såsom att inkorporera flera förträningsuppgifter och ändringar i träningsprocessen, som möjliga framtida forskningsvägar. Sammantaget belyser denna studie de utmaningar och möjligheter som är förknippade med att kombinera förträning med ett diskret flaskhalslager inom kontexten av simulerad flygnavigering och förstärkningsinlärning.