test report

Prepared for Timothy J. Wilmott ’80 Timothy J. Wilmott ’80 Endowed Chair in the Department of Industrial and Systems EngineeringThank you for advancingthe Department of Industrial and Systems Engineering

Thank You for Supporting Exceptional Faculty Endowed faculty positions are essential to Lehigh University because they offer the opportunityto recruit and retain leaders in research and education. Outstanding faculty members are attracted toand remain at the university because their contributions to teaching and research are held in high regard.Building an extensive pool of talented faculty enriches the academic environment, which in turn, attractsthe brightest students. Your support of dynamic faculty through the Timothy J. Wilmott ’80 Endowed Fund is a greatasset to the university as it enables professors to focus on mission-critical initiatives, develop expertisein emerging high-impact areas, and compete more effectively for funding in important cross-disciplinaryfields. Lehigh currently employs nearly 100 endowed faculty members who individually offer substantialcontributions to teaching, scholarship, and research in their chosen disciplines. We are proud of P.C.Rossin College of Engineering and Applied Science professors like Alexander Stolyar, PhD, the Timothy J.Wilmott ’80 Endowed Chair in Industrial and Systems Engineering, who is featured in this report, becausehe provides an educational experience that is of the highest standards through the incorporation of hispersonal industry experience, extensive knowledge, and research. Tim, we are thankful that your positive Lehigh experience as an industrial engineering student,inspired you to support the program so extensively. Your generous contributions to support the renovationsof Mohler Lab as well as faculty through the creation of the Timothy J. Wilmott ’80 Endowed Chair inIndustrial and Systems Engineering, has been instrumental in the success of the industrial and systemsengineering department and we will remain forever grateful for your generosity. 1

Alexander L. Stolyar, PhD - Timothy J. Wilmott ’80 Endowed Chair Alexander L. Stolyar, PhD, the Timothy J.Wilmott ’80 Endowed Chair and Professor in theDepartment of Industrial and Systems Engineeringis researching one of the latest technologicalrevolutions—cloud computing—and his work ofdesigning and studying algorithms focuses on keepingmuch of the computer processes that we take forgranted running efficiently. Since his arrival at Lehighin 2014, the Russian-born Stolyar has been sharingthe knowledge he gained while working at Motorola,AT&T and most recently, Bell Labs, where he was amathematical sciences researcher for 16 years. “The Professor Stolyar assists graduate student Mohammadindustry perspective and history that I can communicate Reza Nazari ’19G with his research on dynamicare important when interacting with students,” he matching systems. Nazari ‘s research interestsexplains. “Because the problems must always come include applied probability, queueing systems, andfrom something real.” To underscore this practicality, optimization.Stolyar points out that the “father” of queueing theory, a major focus of Stolyar’s research, was Agner Erlang, aDanish engineer who worked for the Copenhagen Telephone Exchange 100 years ago. “Erlang was merly tryingto determine how many telephone lines were needed so people wouldn’t have to wait long to place a call,”explains Stolyar. Similarly, Stolyar’s research related to keeping data moving efficiently is just as important tocomputer and cellphones users today. Stolyar is investigating the load distribution in the cloud, which determines the best way to migrate tasks to underutilized virtual machines in order to effectively“Professor Stolyar always tries to convey share resources. Stolyar uses the analogy of a supermarket tohis experiences in industry to his students explain this process, “At the supermarket, there are a numberand this brings great value to his teaching of cashiers, each with a queue, or line of waiting customers.style and research. I’ve had three courses Which line do you choose? Imagine these cashiers are tens ofwith him: stochastic processes, queueing thousands of computers in the cloud. If you want to place yoursystems, and advanced inventory theory. job in the cloud, where should it go? In order to make this decision, we require a distribution algorithm that works wellThese classes were interactive and taught in the sense that large queues will not build up while utilizingat a doctoral level so we could use the the least amount of computational power. We also require anmaterials in our research.” algorithm that works with as little communication as possibleMohammad Reza Nazari ’19G between the computers.” 2

Bringing Industry Experience to Lehigh Students Queueing theory, or the mathematical study of waiting lines or queues, is a concept that has beenof interest to Stolyar since he was young. “By age 12, I was interested in math but even then, I was drawn toprobability and random processes. Somehow I always knew this is what I would be doing.” Stolyar specializesin stochastic processes and their mathematical models, which are used in a wide variety of settings inscience, engineering, finance, and economics to understand events or systems affected by random variables. Stolyar explains that stochastic modeling is crucial to the worldof the cloud and high performance computing. “Cloud computing isoccurring much faster than anyone predicted. A large percentage ofbig data processing is now completed in the cloud, a technology that isimmense and fascinating,” he explains. “These days, when you use yourcomputer, you do not necessarily need the processing power of yourdesktop. You can request a virtual machine in the cloud from companieslike Amazon, for example. However, the virtual machine that you arerequesting has to be “placed” on a large physical computer in the cloud.”Determining the best way to fit all the pieces together into these cloudcomputers is like the mathematical puzzle of packing objects into arectangular box. Stolyar’s research is focused on how best to ‘pack’ thevirtual computers. His research is applicable to many fields outside ofcloud computing including telecommunication networks, inventory management, production systems, andcall centers. “People are very interested because such problems have many other applications,” he says.About Alexander L. Stolyar, PhD Stolyar earned a master’s degree in applied mathematics in 1982 from the Moscow College ofEngineering and a doctorate in mathematics in 1989 from the Institute of Control Sciences. Because hiscountry was in political turmoil in the early 90s, he left Russia and moved to the U.S. Stolyar’s first positionwas with a research group at Motorola and four years later, he was hired at AT&T. In 1998, he accepted a“dream job” with Bell Labs. “A great deal of “firsts” occurred at Bell Labs,” Stolyar explains. “The cell phone,C/C++ computer programming languages, UNIX operating systems, and information theory, were all inventedthere.” 3

Research and TeachingResearch InterestsDr. Stoylar’s research focuses on communication, information andservice systems including scheduling, routing, congestion control,and quality of service. He is also interested in stochastic dynamicbin packing, dynamic matching models, large scale data centers,call/contact centers, systems with on-demand servers, and networkcloud data centers. Stochastic queueing networks also remain aresearch priority for Dr. Stolyar, including his interest in dynamiccontrol and optimization, asymptotic methods, stochastic stability,fluid and diffusion limits, and scaling properties. Stolyar is an associate editor of Queueing Systems andAdvances in (Journal of) Applied Probability. He holds 14 UnitedStates patents and has routinely submitted patent applicationsannually for systems and methods that he has developed relatedto scheduling and allocating data and resources pertaining tonetworks.TeachingDr. Stolyar currently teaches Queueing Systems, a 400 level class that is part of the curriculum for theMaster of Engineering (M.Eng.) degree in the Management Science and Engineering program. Theprogram is directed toward integrating scientific methods with the functional aspects of organizations byinvestigating the application of quantitative methodology and systems analysis in the context of decisionmaking, risk analysis, economics and cost analysis, production management, and supply chain logistics.This integration provides students with a broader perspective toward managerial decision making in bothprivate enterprise and public administration.Selected PublicationsA.L. Stolyar, Large-scale heterogeneous service systems with general packing constraints, Advances inApplied Probability, 2017, Vol. 49, No.1.G. Pang, A.L. Stolyar, A service system with on-demand agent invitations, Queueing Systems, 2016, Vol. 82,No. 3, pp. 259-283.L. Nguyen, A.L. Stolyar, A service system with randomly behaving on-demand agents, SIGMETRICS-2016. 4

Financial Summary Timothy J. Wilmott ’80 Endowed FundData as of Book Value Market Value Spending Distribution6/30/2016 $2,000,000 $1,896,863 for FY 16 $88,248.64Recognizing the need to provide a steady, reasonably predictable stream of income while protectingthe real value of the endowment’s principal, the university has adopted a spending policy distributionrate of 5% based on a twelve-quarter moving average market value. As a further protection againstunpredictable annual fluctuations in the amount of endowment spending distribution, Lehigh also limitsthe year-to-year change for each fund to an amount between 0% and 10%. For the year that endedJune 30, 2016, the total spending distribution from the endowment provided more than $65 million tothe Lehigh budget to benefit university students, faculty, and programs. Lehigh University EndowmentPerformance is summarized below. Perform ance as of 6/3 0/2016 20 15-16 3-Yea r 5 -Year 10-Yea r Lehigh U niversi ty Endo wment 2 .4% 4.5% 4.1% 3.7%LU Composite Benchmark* 1.5% 4.2% 3.4% 3.4% * Lehigh University’s composite benchmark is a blended benchmark consisting of indices for all asset classes. 5