Some Recent Results on Linear Programming Based Approximate Dynamic Programming
The linear programming based approximate dynamic programming has received considerable attention in the recent literature. In this approach, high dimensional dynamic programs are solved approximately as large-scale linear programs to tackle the curse of dimensionality. The linear programming formulations are called approximate linear programs (ALPs) and typically have a large number of decision variables and constraints. A major challenge of the approach therefore lies in efficient solution of the ALPs. In this talk, I report some recent applications and theoretical results in this area of research. Example applications include network revenue management, medical appointment scheduling, and queueing control. I will conclude with discussions on research directions and potential applications in other application areas.