Srinivas Chakravarthy, Ph.D. - US grants

9313283 Chakravarthy Finite capacity queuing systems arise naturally in many practical situations, notably in communications, manufacturing and production systems. These queuing systems have very few analytical results (most of which are obtained under very stringent conditions such as exponential assumptions) and have been getting a lot of attention in recent years due to their profound applications in practice. In this research using a dynamic group service rule that has a number of potential application, several finite capacity queuing models are investigated. Using matrix analytic methods, these models will be analyzed. Efficient and stable algorithms for computing various system performance measures will be developed. These algorithms are to be used interactively to solve non-traditional but significant design problems. An attempt is made to narrow the gap between theoreticians and practitioners in operations research and production system areas which will be beneficial to industry.

Chakravarthy This project is aimed at bringing together in a conference forum researchers, practitioners, and students with interest in matrix-analytic methods for modeling stochastic systems. The conference has multiple objectives. These objectives include 1) review of the progress of research and the state-of-the-art in matrix analytical methods in stochastic models, 2) the analysis of the design of a tutorial that would allow providing a working knowledge of matrix-analytical methods to practitioners and students in a form that they can put the technique to use, 3) discuss and identify technological areas of application of the technique, 4) present resent unpublished research results, and 5) identify directions for future research. Matrix-analytical method for modeling stochastic systems is a relative new area with the first set of systematic studies appearing in the 1980s. Since then, a number of researchers in various fields such as telecommunication, production, and computer systems have applied it to model problems. Despite their widespread areas of applications, matrix-analytical procedure is still relatively unknown to many potential users and researchers in the general area of stochastic analysis. The conference will help to close the gap and disseminate the technique and its potential application to a much wider audience than has been so far possible.

The proposed conference titled "Flint: One City - One Hundred Years Under Variability" is scheduled to be held at Kettering University, Flint, Michigan, June 24-28, 2014. One of the main goals of the project is to bring together experts in the fields of predictive statistics with expertise in statistical modeling, education, and applications. A specific emphasis is on statistics of historical data. Through this event, the organizers intend to celebrate the International Year of Statistics and the 175th anniversary of the American Statistical Association.

The practitioners in engineering and applied sciences are now extensively using statistical methods to solve problems arising in real-life situations dealing with local and federal governments, medical services, insurance, actuaries and public media. While most theoretical aspects and the range of applications of statistical methods have been well established, the organizers of this conference plan to bring together the most active and best researchers, educators, and practitioners to a common setting. The organizers hope that the participants such as students, teachers, and practicing statisticians will take back their experience to enhance their careers as well as to promote the use of statistics in day-to-day activities to benefit society.

The primary objective of the SAXS program at BioCAT is to provide the biomedical community access to the small angle x-ray scattering instrument and to maintain the state-of-the-art. BioCAT has nurtured a user program with a rich history of productive collaborations that include a diverse set of biomedical interests. User support is provided 24/7 for onsite users, and BioCAT provides remote collaborations that involve a high degree of beamline personnel participation. There are two broad categories of experiments performed under the SAXS program: equilibrium and time-resolved SAXS. Liquid chromatography SAXS (LC-SAXS) constitutes > 90% of the equilibrium SAXS program. LC-SAXS in combination with a coflow sample cell, based on a design first adopted by the Australian SAXS beamline, has streamlined SAXS data acquisition at BioCAT by nullifying common concerns such as sample heterogeneity and radiation damage induced capillary fouling. Introduction of in-line multi-angle light scattering (MALS) and dynamic light scattering (DLS) detectors has further enhanced the scope and quality of data acquired using our equilibrium SAXS instrument. As part of the time-resolved SAXS program, three different options are available which cater to time regimes from ~80 microseconds to several seconds. The stopped-flow mixer or a laminar-flow mixer are recommended for accessing time regimes between a millisecond and a few seconds, depending on the amount of sample available and the system being studied. The chaotic flow mixer is used to access earlier time regimes (~ 80µs to ~ 90 ms). Optimized fluid delivery systems, coordinated triggering mechanisms, and GUI based user-friendly software tools developed in house have now made the continuous flow time-resolved SAXS instrument at BioCAT one of its kind in the world and open to the general user community. BioCAT also maintains the widely used open-source data analysis program BioXTAS RAW. Maintenance and upgrade plans cover improved liquid chromatography systems, data analysis software, and microfluidic mixers, as well as upgraded computing, data storage capabilities, network infrastructure, and detectors. We will fully optimize the instrument to take advantage of the world class beam that will be provided by the APS-U project. Last but not the least we will continue to increase automation for SAXS data acquisition in order to make the user as independent as possible.