office: NSH 3211
address: 5000 Forbes Ave
Pittsburgh, PA 15213
phone: 412 268 5581
Scalable planning under uncertainty with repeated graph searches
These projects aim at developing algorithms that can solve large-scale planning under uncertainty problems by utilizing graph searches. Graph searches are general, efficient and come with rigorous theoretical analysis. They typically do not apply however to solving problems that involve uncertainty (i.e., Markov Decision Processes). In these projects, we study how various planning under uncertainty problems can be decomposed into a series of graph searches and thus solved efficiently and with solid theory.
Here are few examples of our past work: