Yildiray Yildiz gave a talk titled “A Game Theoric Framework for Predicting the Evolution of Complex Systems with Human Elements” at the University of Nevada at Reno on September 20, 2013. Human decision-makers play a major role in the operation of most real-world systems of today. The successful operation of these systems often hinges upon the sound judgment of a few individuals. For example, pilots and air traffic controllers continuously make decisions that determine the safety and operation of the National Airspace System. Even if replacing the humans with automation is conceivable, it will be many decades before dependence on human decision-making becomes negligible.
Since humans play such a crucial role in characterizing real-world systems, it follows that to make any accurate predictions about system behavior requires a model that is capable of capturing both the human and non-human dynamics of the system. In this talk, Yildiray presented a game-theoric framework for predicting the evolution of complex systems with human elements, and described how this framework is used to predict human decisions in midair aircraft conflicts, aircraft merging and landing, and cyber-attacks on smart grids.
BACKGROUND: Yildiray Yildiz is an associate scientist for the Adaptive Control and Evolvable Systems group in the Autonomous Systems and Robotics research area. He is currently working on a project to predict pilot reactions using game theory and reinforcement learning. Past projects include work on a framework for telerobotics across the time delays of space, air-breathing hypersonic vehicle control, pilot-induced oscillation recovery using optimal control allocation, and stable adaptive control of automotive powertrains (air fuel ratio control and idle speed control).
NASA PROGRAM FUNDING: The Verification and Validation of Flight Critical Systems (VVFCS) element of the System-wide Safety and Assurance Technologies (SSAT) project, Aviation Safety Program (ASAP)
Contact: Yildiray Yildiz