Benjamin Melamed (PhD, University of Michigan, 1976) is a Distinguished Professor and a former Senior Associate Dean for Strategic Planning and Implementation—New Brunswick. His research interests include supply chain management, supply chain financial management, and service chain management (including modeling, analysis, simulation and optimization), general systems modeling and performance evaluation, stochastic processes, traditional and hybrid simulation (discrete-event and fluid-flow paradigms), and decision support tools.
Melamed authored or co-authored over 100 papers and co-authored two books, and has published in a broad range of scientific journals, including Operations Research, Mathematics of Operations Research, Management Science, J. of Applied Probability, Advances in Applied Probability, J. of Stochastic Processes and their Applications, IEEE Trans. on Automatic Control, Annals of Operations Research, Stochastic Models, Journal of Business Logistics, Performance Evaluation, J. of Optimization Theory and Applications, Nonlinear Analysis, JACM and QUESTA.
Melamed became AT&T Fellow in 1988 and IEEE Fellow in 1994.
Ph.D., University of Michigan; Computer and Communications Sciences
Ph.D. Research Areas
My current broad research area with PhD students is supply chain management. This work calls for students with strong analytical and computer skills. I am interested in modeling, simulation and analysis of production-inventory systems, including the following aspects:
- Discrete and fluid-flow networks production-inventory systems. Here my focus is on optimization of performance metrics using IPA (Infinitesimal Perturbation Analysis) derivatives, since IPA in fluid models gives rise to unbiased nonparametric gradient estimates. Applications include design and control of production-inventory systems, manufacturing lines, bulk material handling, and telecommunications networks.
- Discounted costs in production-inventory systems. Traditional optimization methods use ordinary costs to obtain objective functions and subsequently carry out the corresponding optimization. However, discounted costs (present value of future costs discounted to time 0) are sounder from a financial viewpoint.
- Autocorrelated time series. Realistic models of time series require models that are fitted to a strong statistical signature of empirical data. Typical applications include arrival processes at production-inventory systems, queueing systems, etc. Accordingly, I defined a class of versatile continuous-space, discrete-time stochastic processes, called Autoregressive Modular (ARM), which fits a versatile model to the empirical histogram and autocorrelation function, simultaneously. Recent work aims to extend this approach to a class of multivariate time series, dubbed Multivariate Autoregressive Modular (MARM) processes.
Publications with PhD Students and Alumni
B. Melamed, S. Pan and Y. Wardi, “Hybrid Discrete-Continuous Fluid-Flow Simulation”, Proc. of the SPIE International Symposium on Information Technologies and Communications (ITCOM 01), Scalability and Traffic Control in IP Networks, 263--270, Denver, Colorado, August 22-24, 2001.
B. Melamed, S. Pan and Y. Wardi, “HNS: A Streamlined Hybrid Network Simulator”, ACM Transactions on Modeling and Computer Simulation (TOMACS), Vol. 14, No. 3, 1-27, 2004.
B. Melamed and S. Singh, “Parallelization Algorithms for Modeling ARM Processes”, J. of Applied Mathematics and Stochastic Analysis, Vol. 13, No. 4, 393--410, 2000.
D. Jagerman, A. Altiok, B. Melamed, and B. Balcioglu, “Mean Waiting Time Approximations in the G/G/1 Queue”, QUESTA, Vol. 46, 481-506, 2004.
Name: S. Pan
Rutgers Center for Operations Research (RUTCOR)
Graduation Date: 2005
Thesis Title: "Hybrid Network Simulation"
Name: Shi, Junmin
Rutgers Business School, Supply Chain Management
Graduation Date: 2010/October
Thesis Title: Make-to-Stock Production-Inventory Systems with Compound Poisson Demands, Constant Continuous Replenishment and Lost Sales.
Postdoctoral Research Supervised:
S. Singh, "Parallelization Algorithms for Modeling QTES Processes", postdoc supervisor, Rutgers Business School, Rutgers University, 1998-2000.
Early Summer Research Projects of Current PhD Students:
Name: Dinesh Pai
Project Title: The Impact of RFID on the Security and Integrity of the US Pharmaceutical Supply Chain
The complexity of US pharmaceutical supply chains is increasing rapidly. Demographic changes, growing Internet pharmacies, counterfeit drugs, product diversion and the issues of drug importation and re-importation have added to this complexity. Growing cases of counterfeit drugs and product diversion have posed a serious challenge to drug manufacturers as well as other supply chain partners striving to ensure that the end patient receives authentic product. These have served to raise American awareness of the problem.
The FDA Counterfeit Drug Task Force has recommended a combination of rapidly improving track-and-trace and product authentication technologies to protect the US pharmaceutical supply chain. RFID technology, combined with recent AutoID initiatives led by MIT, is gaining momentum. RFID provides visibility of products along the supply chain, and accurate and timely information, which will help detect counterfeit drugs in the supply chain.
This paper overviews RFID technology and describes how RFID technology could help improve the security and integrity of the drug supply chain. A special section focuses on the concerns of drug manufacturers as well as other supply chain partners regarding the implementation of this technology.
Yihong Fang: IPA derivatives is make-to-stock systems
Junmin Shi: ARM (Autoregressive Modular) process forecasting of financial time series.