OORDSU: from offline HPC to online embedded decisions

I am organizing a growing part of my research around OORDSU: Offline-HPC to Online-Embedded Real-time Data-driven Sequential decision making under Uncertainty.

The basic idea is to treat supercomputing, embedded control, adaptation, and uncertainty as one computational loop rather than separate research topics. Expensive offline computation should produce representations and policies that survive contact with real-time hardware; online experience should in turn inform what the offline computation does next.

See the OORDSU research page →