Font Size:

Technical Program

March 3, 2008 - Hotel Mercure Marseille Euro Centre

Top

Keynote Speech

Hybrid Stochastic Fluid Simulation: Modeling, implementation and IPA

by Prof. Benjamin Melamed
Rutgers University - Rutgers Business School
Newark and New Brunswick, Department of Management Science & Information Systems
Piscataway, NJ 08854 USA

ABSTRACT

Fluid-flow models can be used to complement traditional queueing systems as alternative models of workload transfer, with applications to telecommunications networks, bulk material transport and handling, and more recently, make-to-stock production/inventory systems. In particular, traditional queueing models of telecommunications networks with packet traffic often lead to infeasible simulation models due to the excessive number of events that need to be processed. In this talk, we describe a simple class of fluid-flow models, called Stochastic Flow Models (SFM), which possesses several important advantages over discrete queueing models: (1) SFMs often have simpler dynamics leading to simpler sample path representation; (2) SFMs can often be simulated at a fraction of the computational effort and memory called for by their discrete queueing counterparts; and (3) unlike traditional queueing systems, SFMs admit unbiased and nonparametric Infinitesimal Perturbation Analysis (IPA) gradient estimators of low computational complexity, so that IPA gradients can be computed either from a simulation run or from observations of a real-life system. Consequently, SFMs can be used in performance evaluation and prediction via discrete-event simulation or real-life systems, and their nonparametric nature holds out the promise of applications to network design, provisioning and control of networks that model workload flow.
The talk will discuss three topics relating to the SFM class and its applications to modeling and simulation of telecommunications networks. First, we define a basic SFM class to serve as building blocks of SFM networks. Second, we review key implementation issues of hybrid simulators admitting both packet streams and fluid-flow streams, highlighting some peculiar aspects of fluid traffic. These are based on our prototype simulator, dubbed Hybrid Network Simulator (HNS), and include data structures and algorithms for unified handling of hybrid workload, workload fragmentation, computational complexity, and protocol mapping. Finally, we describe workload-related performance measures, such as time averages of loss volume and buffer contents, and exhibit their IPA derivatives with respect to various design parameters, such as buffer size, arrival rate and service rate, all in SFM setting (IPA-over-SFM). The talk will conclude with remarks on the potential applications of IPA-over-SFM as an ingredient in optimizing the design, configuration, and possibly control of high-speed telecommunications networks.

Biography of the keynote speaker

Benjamin Melamed is a Professor II at the Rutgers Business School – Newark and New Brunswick, Department of Management Science and Information Systems, the co-director of the DIMACS-CAIT Laboratory for Port Security (LPS) at Rutgers University, a Research Fellow at the Center for Supply Chain Management (CSCM), and a member of the Rutgers Center for Operations Research (RUTCOR) and the Center for Discrete Mathematics and Theoretical Computer Science (DIMACS). Melamed received a B.Sc. degree in Mathematics and Statistics from Tel Aviv University in 1972, and a M.S. and Ph.D. degrees in Computer Science from the University of Michigan in 1973 and 1976, respectively. From 1977 to 1981 he taught at the department of Industrial Engineering and Management Science at Northwestern University. He joined the Performance Analysis Department at Bell Laboratories in 1981, and later became an AT&T Fellow. Melamed moved to NEC in 1989 and served there as a Deputy Director, Head of the Performance Analysis Department, and NEC Fellow. He consulted at Bellcore in 1995 and joined Rutgers University in 1996. Melamed’s research interests include system modeling and analysis (especially supply chains and telecommunications systems), simulation, stochastic processes and visual modeling environments. He authored or co-authored over 100 papers and co-authored two books: “Modern Modeling and Simulation” (with R. Rubinstein), John Wiley and Sons, 1998, and “Simulation Modeling and Analysis with Arena” (with T. Altiok), Academic Press, 2007. His research has been supported by DARPA and NSF. Melamed was awarded an AT&T Fellow in 1988 and an IEEE Fellow in 1994. He became an IFIP WG7.3 member in 1997, and was elected to Beta Gamma Sigma in 1998.

Top

Paper Abstracts

Top