Ward Whitt

Ward Whitt

Past Awards

M&SOM Best Paper Award : Finalist
Winning material: Are Call Center and Hospital Arrivals Well Modeled by Nonhomogeneous Poisson Processes? (2014) M&SOM 16 (3):464-480

MSOM Distinguished Fellows Award : Awardee(s)

Saul Gass Expository Writing Award: Winner(s)

It is with great pleasure that the INFORMS Committee for the Expository Writing Award names Professor Ward Whitt of Columbia University as the recipient of its 2011 award. Ward Whitt is renowned for the many theoretical and applied results he has developed over 40+ years. As the author or coauthor of more than 325 journal articles in addition to his book Stochastic-Process Limits (Springer, 2002), Ward Whitt is a master expositor. His writing is exceptionally clear, and consequently he has succeeded in communicating his many findings to a large audience.

Examples of Professor Whitt’s expository skill are found in two of his early papers, “Approximating a Point Process by a Renewal Process” (Operations Research, 1982) and “The Queueing Network Analyzer” (Bell System Technical Journal, 1983). The first of this pair develops the theory underlying approximations for point processes that can be used to describe non-Markovian traffic in queueing networks or multiechelon inventory systems, while the second reports a software tool employing these approximations to analyze queueing networks with complicated topologies that allow customer traffic to merge, split or feed back as it flows through the system. These papers have each been cited hundreds of times, illustrating their import to the development of the literature in this area.

Ward Whitt’s search for practical tools with rigorous foundation is further demonstrated in his paper with Shlomo Halfin titled “Heavy-Traffic Limits for Queues with Many Exponential Servers” (Operations Research, 1981). Also heavily cited, this article established a new heavy traffic limit with a delay probability less than 1; the resulting Halfin-Whitt regime (often referred to as the QED or quality-and-efficiency-driven regime) now features regularly in analyses of many-server call centers and other service networks. Professor Whitt’s career-long interest in developing practical results for difficult queueing problems continues to the present day; indeed one of his most recent papers, “A Network of Time-Varying Many-Server Fluid Queues with Customer Abandonment (Operations Research, 2011, coauthored with Yunan Liu) testifies further to his clear and careful writing on these topics.

Professor Whitt’s expository output is even more impressive when one considers the journals in which he writes. Research reported by Ward Whitt and his coauthors appears regularly in the very best journals in operations research and stochastic processes including Operations Research, Management Science, Mathematics of Operations Research, Journal of Applied Probability, Advances in Applied Probability, Queueing Systems, and other leading outlets. In addition to his many scientific papers, Ward Whitt’s book Stochastic-Process Limits is a 600+ page treatise that brings together Donsker’s functional central limit theorem with queueing processes to create heavy-traffic stochastic-process limits for queues. These limits have the advantage of revealing key relationships and performance details that are often obscured when attempting to model such queues directly. The book is written so that readers with differing backgrounds in probability theory can learn at appropriate levels of technical expertise.

Professor Whitt’s writing is both clever and clear, and much of his material is indeed beautiful. For more than 40 years he has been hugely influential in propelling stochastic operations research forward in both theory and application. For these reasons, the Expository Writing Committee (Warren B. Powell, Richard Steinberg and Edward H. Kaplan) is pleased to name Ward Whitt as the recipient of the 2011 INFORMS Expository Writing Award.

Frederick W. Lanchester Prize: Winner(s)

Professor Whitt is honored for his book, Stochastic Process Limits: An Introduction to Stochastic-Process Limits and Their Applications to Queues, (Springer-Verlag, New York, 2002).

Stochastic process limits have become a standard tool in operations research to study queues arising from complex systems like communication networks. These limits are often obtained by the scaling of time and space when the system is in heavy traffic, a state in which important features of a control policy are displayed in sharpest relief.

Whitt's book is a lucid treatment not only of traditional heavy traffic limits, built from sums of independent random variables with light tails, but also limits with nonstandard scaling and nonstandard limit processes. These new limits are essential to understanding the long-range dependence structures and heavy tailed distributions that have been observed in evolving communication networks. By accounting for features that were viewed until quite recently as "nonstandard" or "esoteric" at best, Whitt's book lays the foundation for the "queueing theory of the future".

INFORMS Elected Fellows: Awardee(s)

John von Neumann Theory Prize: Winner(s)

Ward Whitt of AT&T Labs is awarded the 2001 von Neumann Prize for his exceptionally broad and profound contributions to queueing theory, applied probability and stochastic modeling.

Whitt’s sequence of three pioneering papers in 1970 (with Donald Iglehart) on multiple channel queues in heavy traffic, his 1974 paper on the continuity of queues, and his 1980 paper on functional central limit theorems, along with many related contributions, have created a lasting and vibrant framework for the analysis of queuing systems through asymptotic approximations. Indeed, much of the most outstanding research in stochastic OR of the past 25 years was built on the foundation established by Whitt's groundbreaking works.

Whitt's research has also advanced the field of communication network performance analysis for several decades and through several generations of technology. He has led the field of traffic modeling, combining practical insights with limit theorems to develop fundamental tools.

Throughout his enormously productive research career, Whitt has also made many other seminal contributions, which often helped define an entire field. These include numerical inversion of transforms (with Joseph Abate and others), simulation methodology (with Peter Glynn) and the far- reaching manifestations of Little's Law (also with Peter Glynn).

Frederick W. Lanchester Prize: Honorable Mention

Honorable mention in the 1997 Lanchester Prize competition is awarded to Joseph Abate, Gagan L. Choudhury, Kin K. Leung, David M. Lucantoni and Ward Whitt for a sequence of fifteen papers on numerical techniques for transform inversion, and application of those techniques to various stochastic models. The deep and scholarly work of these authors is concerned with deriving numerical performance estimates from stochastic models of computing and communication systems.

One can often derive mathematical expressions for the transforms of distributions that characterize performance of a stochastic system model, but solutions in the transform domain have long been viewed as unsatisfactory, at least partly because numerical inversion of the transforms was considered impractical. Abate et al have undertaken a sustained effort to show that notion is mistaken, although one must exploit the special character of stochastic system models if numerical inversion is to be rendered feasible. They have taken a general method of transform inversion and shown how it can be specialized and refined to treat several families of models, some of them quite large and complicated, developing routines that are fast and accurate.

Abate and his colleagues also derived a number of new and impressive transform results (that is, formulas for Laplace transforms themselves, as opposed to inversion methods), and they were often able to derive good error bounds for their inversion routines. It is not just one family of models for which fast and accurate numerical methods were developed, but a number of such families — including single-server queues with many input sources, a large class of product-form loss networks, and polling models. Their work won praise from the committee both for its originality and for opening up new domains of application.

Outstanding Simulation Publication Award: Awardee(s)
Winning material: "Planning queueing simulations." Management Science 35 (11): 1 1989.