John N. Tsitsiklis

John N. Tsitsiklis

Past Awards

2018
John von Neumann Theory Prize: Awardee(s)
2018 - Awardee(s)
Citation:

The 2018 INFORMS John von Neumann theory prize is awarded to Dimitri P. Bertsekas and John N. Tsitsiklis for contributions to Parallel and Distributed Computation as well as Neurodynamic Programming.  Working together and independently, Bertsekas and Tsitsiklis have made seminal contributions to both these fields. They unified ideas and built solid theoretical foundations while these fields were still relatively nascent, thus greatly enhancing subsequent development of rigorous theory.
Their monograph Parallel and Distributed Computation: Numerical Methods represents a significant achievement in the field. The work builds on and extends the authors’ extensive previous work in this area, identifying the tolerance of algorithms to asynchronous implementations and a number of positive convergence results. An antecedent work of particular significance to the operations research community is the paper by Tsitsiklis, Bertsekas and Athans which provides seminal analysis of asynchronous implementations of deterministic and stochastic gradient algorithms. This line of inquiry has recently found application in the analysis of descent algorithms for neural network training and other machine learning problems. Their work in distributed computation has also had significant impact on the areas of distributed network control and distributed detection.
Their monograph Neuro-Dynamic Programming helped provide a unified theoretical treatment of the wide variety of reinforcement learning algorithms by building connections to the dynamic programming and distributed computation literature. This has proven extremely valuable in bringing theoretical rigor to a field of rapid, empirical innovation.  The authors’ contributions in this area go beyond providing a theoretical foundation that other could build on. The authors have made significant original contributions to value function learning, temporal difference methods and actor-critic algorithms.
The work of Bertsekas and Tsitsiklis is characterized by its innovation, depth and clarity, and it has had tremendous impact as evident from the large number of citations. Their two joint monographs are among their individual five most cited works, making the award of a joint prize particularly appropriate. Bertsekas and Tsitsiklis have brought the fields of computer science and operations research closer together through unifying theory.



2017
Saul Gass Expository Writing Award: Winner(s)
2017 - Winner(s)


2012
INFORMS Computing Society Prize: First Place
2012 - First Place


2007
INFORMS Elected Fellows: Awardee(s)


1997
INFORMS Computing Society Prize: First Place