Cyrus Derman

Cyrus Derman

Past Awards

John von Neumann Theory Prize: Winner(s)

The 2002 John von Neumann Theory Prize is awarded by the Institute for Operations Research and the Management Sciences to Donald L. Iglehart and Cyrus Derman for their fundamental contributions to performance analysis and optimization of stochastic systems.

In a series of important papers culminating in the book, Finite State Markovian Decision Processes, Derman fundamentally advanced finite-state-and-action Markovian decision processes. He took the lead role in showing that starting from a state, the set of state-action frequencies over all policies is the convex hull of the finite set of state-action frequencies over all stationary deterministic Markov policies. This work plays a fundamental role in solving such problems in the presence of linear constraints on the state-action frequencies, e.g., reflecting desired limits on the frequency of unfavorable events like failures, rejects, shortages and accidents, and has been widely used in practice. Moreover, it led him to co-develop the first general method for finding minimum-average-cost (typically randomized) policies in the presence of such constraints. With Jerome Sacks, he discovered the celebrated and widely used conditions that assure an optimal stopping policy is myopic, and applied them to give simple solutions of problems of optimal equipment replacement and search for a favorable price to buy an asset. With Gerald Lieberman and Sheldon Ross, he introduced and gave a lovely solution of a sequential stochastic assignment problem that has application to optimally assigning workers with differing skills to randomly arriving jobs of varying types. With his student and long-time colleague, Morton Klein, he originated and did pioneering work characterizing when FIFO and LIFO issuing policies are optimal for items whose value or field life depends on their age at issue. He developed a beautiful solution to the problem of finding a minimax inspection schedule for randomly failing equipment. He has often influenced and contributed to engineering practice by, e.g., co-developing blood-inventory management policies that were used subsequently in New York hospitals, evaluating the safety of New York bridges, and co-developing strategies for early termination of tests for the Navy.

Iglehart pioneered and, in subsequent papers with his student Ward Whitt, led the development of diffusion limits and approximations for heavily congested stochastic systems. The importance of these ideas is that they provide tractable limiting processes and readily computable approximations for complex queueing and other stochastic systems for which closed-form or even numerical solutions have proved intractable. His work in this area transformed the field, with literally hundreds of papers subsequently continuing the development of his ideas. The approximations have proved extremely useful in practice, e.g., in telecommunications and manufacturing systems. His work on diffusion approximations and its extensions has been incorporated into important and widely used software packages. Computer simulation of complex stochastic models of computer, communication and manufacturing systems is a widely used technique for evaluating system performance when analytic methods are insufficient. Iglehart introduced and led the development of the regenerative method for stochastic simulation output analysis. This path-breaking work provided the first sound basis for estimating system performance from simulation with prescribed accuracy, has inspired a flood of significant contributions to simulation methodology and has been embodied in many simulation software packages. His important monograph (Regenerative Simulation of Response Times in Networks of Queues) and series of papers with Gerald Shedler developed many novel system performance models based on these ideas. His subsequent work, with his student Peter Glynn, significantly advanced the subject by incorporating techniques such as importance sampling. With Samuel Karlin, he studied the discounted infinite-horizon inventory problem in which product demands depend on a Markovian index of business conditions. He showed that the optimal ordering policy is base-stock and developed a beautiful and novel method of finding the index-dependent base-stock levels.

Derman and Iglehart both took their PhDs in Mathematical Statistics—Derman from Columbia and Iglehart from Stanford—and are Fellows of the Institute of Mathematical Statistics. Also, Derman is a Fellow of the American Statistical Association and Iglehart is a member of the National Academy of Engineering. Both were attracted to Operations Research very early in their careers and held appointments in the field in engineering schools for about four decades, largely in the institutions from which they took their PhDs. Both are excellent teachers who have inspired many students. Indeed, the PhD students of both have also won several significant prizes and honors, and have been prominent in industry, research laboratories and universities.