The
Linear Programming Approach to Approximate Dynamic Programming
In the late 80's, TD-gammon generated great excitement around approximate dynamic programming methods. The ability of approximate dynamic programming to generate a world-class automated backgammon player suggested that this methodology could be a viable approach to control of large-scale, complex systems. However, almost 15 years later, the use of approximate dynamic programming in real-world applications is still far from being a reality.
This talk will explore the opportunity for new research
in retail operations. The past few decades have witnessed many changes in
retailing. Retailers have become larger (e.g., Wal-Mart alone had sales of
$ 217 billion last year), more prolific users of technology (e.g., US retailers
are estimated to spend more than $30 billion/year on information technology),
and arguably better supply chain managers. This talk will start by examining
the state of retail practice and identify areas that are still a challenge.
These areas can benefit substantially from new research in OR. I will also
present recent research on operational execution and an empirical investigation
of inventory productivity.