ACES-Based Testbed and Bayesian Game-Theoretic Framework for Dynamic Airspace Configuration

The key innovation in this effort is the development of algorithms and a framework for automated Dynamic Airspace Configuration (DAC) using a cooperative Bayesian game framework. Given an initial sector plan, we propose an approach for dynamic re-sectorization in which boundaries can be redefined in response to changing demand, weather or other user preferences. Advantages of this approach are that it models the human coordination process, provides a rich domain independent framework for modeling collaboration and a theoretical framework to analyze issues related to convergence, decision-making complexity and stability. The communicative aspects of the game-framework also make it well suited for an agent-based implementation. In this agent-based implementation, each agent represents a player (Sector ATC/TMC) in the air traffic domain. Sector ATC's collaborate with neighboring Sector ATC's within the current sector, and across center boundaries to "collapse", "split" or borrow airspace to optimize traffic flow. The players engage in an "automated collaborative negotiation search" with each other to determine sector geometries that will optimize the overall airspace efficiency. We propose to implement DAC algorithms in Cybele's Decision Support System Infrastructure and demonstrate feasibility using NASA's ACES software.