NASA is developing algorithms and methodologies for efficient air-traffic management. Several researchers have adopted an optimization framework for solving problems such as flight scheduling, route assignment, flight rerouting, nationwide traffic flow management and dynamic airspace configuration. Computational complexity of these problems have led investigators to conclude that in many instances, real-time solutions are computationally infeasible, forcing the use of relaxed versions of the problem to manage computational complexity. The primary objective of this research is to accelerate optimization algorithms that play central roles in NASA's ATM research, by parallel implementation on Graphics Processing Units (GPUs). This paper focuses on one of the optimization problems viz. the nationwide Traffic Flow Management Problem (TFMP) formulated by as a Binary Integer program. The Binary Integer program has a primal block angular structure that renders it amenable to the Dantzig-Wolfe decomposition algorithm. This research effort implemented a Simplex-based Dantzig-Wolfe (DW) decomposition solver on GPUs that exploits both coarse-grain and fine-grain parallelism. The implementation also exploits the sparsity in the problems, to manage both memory requirements and run-times for large-scale optimization problems. The GPU implementation was used to solve a TFM problem with 17,000 aircraft (linear program with 7 million constraints), in 15 seconds. The GPU implementation is 30 times faster than the exact same code running on the CPU. It is also 16 times faster than the NASA's current solution that implements parallel DW decomposition using the GNU Linear Programming Kit (GLPK) on an 8-core computer with hyper-threading.

As a part of NASA's NextGen research effort, the focus area of Airspace Super-Density Operations (ASDO) performs research pertaining to highly efficient operations at the busiest airports and terminal airspaces. It is expected that multiple ASDO concepts will be interacting with one another in a complex stochastic manner. This research effort developed a high-fidelity queuing model of the terminal area suitable for the design and analysis of NextGen ASDO concepts, as well as to perform time-varying stochastic analysis of terminal area operations with regards to schedule and wind uncertainties. A unique aspect of the current approach is the discretization of terminal airspace routes into 3-nmi servers for enforcing separation requirements. The current research effort developed high-fidelity queuing models of the San Francisco International Airport (SFO) terminal airspace, based on published airspace geometry. A discrete-event simulation framework was developed to simulate the temporal evolution of flights in the terminal area. The queuing simulation framework was used in different case studies involving various phenomena in the terminal area such as compression, conflict and delay analysis, runway reconfiguration and variable inter-aircraft separation. In addition to being a useful analysis tool, the proposed simulation framework shows potential as a real time stochastic decision support tool due to its low computational cost.

The ability to rapidly generate traffic predictions is expected to be central for implementing next-generation air traffic management functionality, both on the ground and aboard aircraft. While high-end computers can be used for this purpose, emerging capabilities of computational hardware such as Graphics Processing Units, together with Cloud Computing concepts can be exploited to realize substantial acceleration of trajectory computations at a modest cost increment. This paper discusses the development of a computational appliance for rapid prediction of aircraft trajectories that combines efficient algorithm and software design with emerging high performance computing architectures. The research effort accelerates trajectory predictions through software profiling and tuning, and implements computationally intensive functions on high performance computing architectures such as computing clusters, multi-threaded programming on multi-core computers and Graphics Processing Units. The fastest of these implementations uses a Graphics Processing Unit, which can perform a system-wide 24-hour trajectory prediction for 35,000 aircraft in less than 2.5 seconds. When compared with the baseline trajectory prediction software, the present approach provides over two orders of magnitude speedup. Nomenclature