by Richard Karp
In areas ranging from earthquake early warning and nuclear event detection to the control of energy grids and urban transportation systems, it is necessary to make real-time decisions based on streaming data arriving in high volume, without human cognition in the loop. A planning workshop on Real-Time Decision Making held at the Simons Institute this summer brought experts in several domains of physical sciences, engineering and societal systems together with mathematical and computational scientists to explore the common features of these decision-theoretic problems and survey the tools required to address them.
Expository talks covered the following application domains: physics at the Large Hadron Collider (Harvey Newman, Caltech); earthquake early warning (Richard Allen, UC Berkeley); semi-automated detection of supernovae (Peter Nugent, LBNL); control of urban bus and train systems (Balaji Prabhakar, Stanford); estimation and control of urban traffic (Pravin Varaiya, UC Berkeley); urban bike sharing (David Shmoys, Cornell); ridesharing platforms (Siddhartha Banerjee, Cornell), air traffic control (Claire Tomlin, UC Berkeley and Dan Feldman, University of Haifa); optimization and control of power networks (Steven Low, Caltech and Sean Meyn, University of Florida); cyber-physical systems (Edward Lee, UC Berkeley), and online advertising (Amin Saberi, Stanford). These talks were complemented by fifteen talks on relevant algorithmic and mathematical topics including convex optimization, control theory, streaming algorithms, computational learning theory, clustering, coding for interactive communication, local computation, market dynamics, dynamic matching models, and risk mitigation in networks.
The themes of this planning workshop will be explored more fully in a semester-long Simons Institute program on Real-Time Decision Making in Spring 2018. The organizers of the workshop and the program are Joshua Bloom (UC Berkeley), Richard Karp (UC Berkeley), Steven Low (Caltech), Evdokia Nikolova (UT Austin), and Balaji Prabhakar (Stanford).