Approximate Dynamic Programming

In spite of considerable research efforts, curse of dimensionality which affects the scalability of approximate dynamic programming (ADP) models still remains a challenge. This work advances the state-of-the-art in the field of stochastic dynamic programming methods through the innovative integration of a diffusion wavelet based value function approximation method with ADP. The innovation lies in this integration that exploits the structure of the problem to achieve computational feasibility. The research method is tested on the problem of taxi-out time estimation of aircraft to establish a proof of concept. The sequential predictions of taxi-out time obtained in real-time for departing aircraft provide shared situational awareness to benefit airport operations planning. The outcomes of this work provide a generic methodology for sequential decision making under uncertainty in large scale applications by uniting concepts from signal processing, statistics, stochastic processes, and artificial intelligence, which may provide solutions for future automated decision making in large scale complex applications in other engineering domains.

Autorentext

Poornima Balakrishna received her PhD in Operations Research from George Mason University (GMU), Fairfax VA. She is affiliated with the Center for Air Transportation Systems Research at GMU and is currently a research engineer at The Sensis Corporation, Reston VA. Her work has been extensively peer-reviewed at leading conferences and journals.

Klappentext

In spite of considerable research efforts, 'curse of dimensionality' which affects the scalability of approximate dynamic programming (ADP) models still remains a challenge. This work advances the state-of-the-art in the field of stochastic dynamic programming methods through the innovative integration of a diffusion wavelet based value function approximation method with ADP. The innovation lies in this integration that exploits the structure of the problem to achieve computational feasibility. The research method is tested on the problem of taxi-out time estimation of aircraft to establish a proof of concept. The sequential predictions of taxi-out time obtained in real-time for departing aircraft provide shared situational awareness to benefit airport operations planning. The outcomes of this work provide a generic methodology for sequential decision making under uncertainty in large scale applications by uniting concepts from signal processing, statistics, stochastic processes, and artificial intelligence, which may provide solutions for future automated decision making in large scale complex applications in other engineering domains.