Bahar Biller
Assistant Professor of Manufacturing and Operations Management   
5000 Forbes Avenue, Posner Hall 360
Pittsburgh, PA 15213
Phone: 412 - 268 3977
mailto:billerb@andrew.cmu.edu
 
 

Curriculum Vitae


Page Map

  1. Education
  2. Research Interests
  3. Professional Experience
  4. Teaching Portfolio
  5. Publications
  6. Working Papers and Software
  7. Presentations
  8. Awards and Honors
  9. Professional Memberships
  10. Professional Activities and Service

Education


Research Interests

My research centers on developing models of stochastic systems and the major application areas are production systems and supply chain. The motivation of my research is that dependent time-series input processes occur naturally in the simulation of many service, communications, and manufacturing systems. For example, the demand for a product during this time period is probably correlated with the demand with the previous and the following periods since economic conditions are most likely similar. Another example is that customers who log on to a particular web site have characteristics that influence their behavior, including age, sex, income level and where they live. To treat these customer characteristics as independent random variables misses the obvious relationship between age and income, for instance. Although ignoring such dependencies can lead to performance measures that are seriously in error and a significant distortion of the simulated system, the common practice is to ignore them! My objective is to quantify the impact of autocorrelation on the system performance and then develop models considering such dependencies. At the moment, I have been working on two different projects:

  1. Quantification of the impact of autocorrelation on the performance measures of various production systems . For instance, if there is actually positive autocorrelation between the interarrival times of customers to a queue but you ignore it, then the simulation of the queue can grossly underestimate the congestion that will actually occur. 
  2. Providing managerial insights to international firms on making better flexible capacity planning . Today, it is well known that the major benefits of flexibility, i.e., the ability to build different types of products in the same manufacturing plant, are increased sales and capacity utilization as a result of hedging against uncertainty. I am interested in considering this problem in an international setting, i.e., manufacturing facilities are located in different countries and exposed to different economic conditions and product regulations (e.g., the emission requirements for automobiles are different in Europe and USA). My objective is to quantify the benefits of flexibility in such a setting and provide managerial insights on making better flexible capacity investment decisions. This research is supported by the Carnegie Bosch Institute

I am also interested in multivariate input modeling . Typical examples from manufacturing and service applications include the sales of different products in different geographic regions over time, the processing times of work-pieces across several work-centers, the times between arrivals of calls to a call center, the annual return on stock and bond investments in a portfolio, and the medical characteristics of organ-transplant donors and recipients. Input modeling -- choosing an appropriate mathematical representation of these phenomena -- is a challenge in today’s input-modeling software packages due to the lack of robust procedures to model multivariate dependencies in the processes. My objective is to fit flexible input models to multivariate non-normal time-series input processes. If successful, the product of this research will be far superior to prior stochastic input models because of its ability to characterize dependencies both in time sequence and with respect to other components in the input process while representing key features of the data at hand through a flexible system of distributions.


 Professional Experience


Teaching Portfolio

Teaching

Development


Publications

B. Biller and B. L. Nelson. 2002. “Modeling and generating multivariate time series with arbitrary marginals using a vector autoregressive technique.” Forthcoming, ACM Transactions on Modeling and Computer Simulation. Available Online Companion.

B. Deler and B. L. Nelson. 2001. “Tutorial: Answers to the top ten input modeling questions.” To appear in Proceedings of the 2002 Winter Simulation Conference. Piscataway, New Jersey: Institute of Electrical and Electronics Engineers.

B. Biller and B. L. Nelson. 2002. “Parameter estimation for ARTA processes.” To appear in Proceedings of the 2002 Winter Simulation Conference. Piscataway, New Jersey: Institute of Electrical and Electronics Engineers.

B. Deler and B. L. Nelson. 2001. “Modeling and generating multivariate time series with arbitrary marginals and autocorrelation structures.” To appear in Proceedings of the 2001 Winter Simulation Conference. Piscataway, New Jersey: Institute of Electrical and Electronics Engineers.

Dincer, C. and B. Deler. 2000. “On the distribution of throughput of transfer lines.” Journal of the Operational Research Society. Volume 11, No. 10, pp. 1170 – 1178.

B. Deler and B. Tansel. 1998. “On the flow-path design formulation for automated guided vehicle systems.” Technical Report IEOR – 9802. Industrial Engineering Department, Bilkent University.

B. Deler and C. Dincer. 1997. “Transient analysis of serial line production systems: A literature review.” Technical Report IEOR – 9712. Industrial Engineering Department, Bilkent University.


Working Papers and Software

B. Biller and B. L. Nelson. 2002. “Fitting time-series input processes for simulation.”. In preparation.

B. Biller and B L. Nelson. 2002. “On the performance of the ARTA fitting algorithm for stochastic simulation.” In preparation.

B. Biller and B. L. Nelson. 2002. “A vector autoregressive framework for time series input processes: Software and applications.” Work in progress.

Click here to download ARTAFIT software for fitting univariate time-series input processes.

Click here to download VARTA software for simulating multivariate time-series input processes.


Presentations

B, Biller and B. L. Nelson. 12/02. “Tutorial: Answers to the top ten input modeling questions .” To be presented at Winter Simulation Conference, San Diego, CA. You can download the ARENA demo files on Autocorrelated Arrival Process and Nonstationary Poisson Arrival Process

B, Biller and B. L. Nelson. 12/02. “Parameter estimation for ARTA processes .” To be presented at Winter Simulation Conference, San Diego, CA.

B, Deler and B. L. Nelson. 12/01. “Modeling and generating multivariate time series with arbitrary marginals and autocorrelation structures.” Winter Simulation Conference, Arlington, VA.

B. Deler and B. L. Nelson. 11/01. “Improving simulation input models through VARTA.” Informs Conference, Miami Beach.

B. Deler and B. L. Nelson. 7/01. “Towards a comprehensive stochastic input-modeling framework for continuous distributions.” INFORMS Applied Probability Society Conference, New York, NY.

B. Deler and B. L. Nelson. 11/01. “Vector autoregressive to anything input processes for simulation.” INFORMS Conference, San Antonio, TX.

B. Deler and C. Dincer. 7/98. “On the distribution of throughput of transfer lines.” INFORMS International Conference, Tel Aviv, Israel.

B. Deler and C. Dincer. 6/98. “On the performance of transfer lines.” 20th National Congress of Operations Research and Industrial Engineering, Ankara, Turkey.

Dincer, C. and B. Deler. 5/97. “On the variance of three-stage two-buffer transfer lines.” INFORMS Conference, San Diego, CA.


Awards and Honors


Professional Memberships


Professional Activities and Service


Last update 11/1/02