Critical Infrastructure Modeling
Organizer: Sallie Keller-McNulty
(sallie@lanl.gov)
Los Alamos National Laboratory

Description:
From the highest levels of government and the private sector to system designers, operators, and scientists, simulations are used to combine data, knowledge, and situational context. These simulations assist in human reasoning and decision-making in complex socio-technical environments such as critical infrastructure systems. Simulation technologies being developed today comprise extremely large complex systems, frequently with hundreds of millions of interacting components (e.g., human interactions). As representations of the real world, these simulations are unique in their ability to illuminate the inner workings and predict the consequences of extremely complex systems.

This session will discuss critical infrastructure models, from their mathematical structure to their predictive capabilities. The models presented here are sequential dynamical systems that incorporate integrated event-based simulation methodologies, unifying multiple problem domains with varying levels of component aggregation and disparate time scales. The results allow for the study of issues related to the characterization of system complexity, sensitivity, robustness, and degree of interdependence, with the identification of critical components, synergistic behavior, and the fundamental limits on the systems' predictive capabilities.

Format:
The session will include three presentations. Darrell Morgeson will present various application areas making use of these models. Chris Barrett will discuss the mathematical underpinnings of the models (i.e., sequential dynamical system). Richard Beckman and Kathy Campbell will discuss the complexity and challenges for the statistical analyses of these simulation models.

Participants:
J. Darrell Morgeson (presentation, Simulation-Based Analysis of the Nation's Critical Infrastructure)
J. Darrell Morgeson is Division Director for the Technology and Safety Assessment (TSA) Division at the Los Alamos National Laboratory. He received a B.S. degree in Engineering from the United States Military Academy at West Point, New York, and a M.S. degree in Operations Research from the Naval Postgraduate School in Monterey, California. From 1971--1984, Darrell was with the U.S. Army, Field Artillery, where his work involved operational systems, 3 years command, advanced weapons concepts, and tactics development. In 1984, he joined the Los Alamos National Laboratory where he has been instrumental in developing the Laboratory's core capabilities in simulations of large-scale, complex, adaptive systems. His current programmatic emphasis is on modeling, simulation and analysis of nuclear systems (e.g., weapons, materials, and energy), non-nuclear defense systems, and critical infrastructure (e.g., transportation, electricity, etc.).

Chris Barrett (presentation, Generation and Measurement of Large Dynamical Systems)
Chris Barrett is Group Leader for the Technology and Safety Assessment Simulation Applications at the Los Alamos National Laboratory. He received a Ph.D. from the California Institute on Technology in Engineering Science/Bioinformationtion Systems in 1985. During the last 10 years, he has been doing theoretical applied research in areas of intelligent control, simulation, and nonlinear dynamics.

Katherine Campbell and Richard Beckman (presentation, Statistics for Complex Computer Models: Beyond Input-Output Analysis)
Katherine Campbell is a Staff Member in the Geoanalysis Group at the Los Alamos National Laboratory. She received a B.A. in physics and M.A. and Ph.D. degrees in mathematics. At EG&G in Los Alamos from 1972 to 1975 and since joining the Los Alamos National Laboratory in 1975, she has worked on statistical applications in environmental investigations and modeling as well as signal and image processing.

Richard Beckman is a Staff Member in the Statistical Sciences Group at the Los Alamos National Laboratory. He obtained a B.S. in mathematics in 1974 and a Ph.D. in statistics in 1969 from Kansas State University. He served on the Biology and Mathematics faculty at the University of Cincinnati from 1969--1971. After joining the Los Alamos National Laboratory in 1971, he has held various management and technical staff positions. Currently, he is one of the leaders of the transportation simulation project, TRANSIMS . He has a long history of publications on outlier detection, sensitivity and uncertainty analysis, and the analysis of computer codes.

Matthias Schonlau (poster, Cyber Crimes and Counter Measures)
Matt Schonlau graduated in 1997 from the University of Waterloo with a dissertation about "Computer Experiments and Global Optimization" under the supervision of William J. Welch. From 1997 through 1999 he worked on computer intrusion detection while pursuing a postdoctoral position with the National Institute of Statistical Sciences and AT&T Labs-Research. In the fall of 1999 Matt joined the RAND corporation. His current projects include estimating the cost of clinical trials, and investigating alternative recruitment policies for the Air Force.

Jeffrey L. Solka, David J. Marchette, and Bradley Wallet (poster, Statistical Visualization Methods in Intrusion Detection)
Jeffrey L. Solka earned the B.S. degree in Mathematics and Chemistry from James Madison University in 1978, the M.S. in Mathematics from James Madison University in 1981, the M.S. in Physics from Virginia Polytechnic Institute and State University in 1989 and his Ph.D. in Computational Sciences and Informatics (Computational Statistics) at George Mason University, working under the direction of Prof. Edward J. Wegman, in May of 1995. Since 1984, Dr. Solka has been working in nonparametric estimation and statistical pattern recognition for the Naval Surface Warfare Center, Dahlgren, Virginia. He has published over 100 journal, conference, and technical papers, has won numerous awards, holds 4 patents, and currently holds a SECRET security clearance.

David J. Marchette received a B.A. in 1980, and an M.A. in mathematics in 1982, from the University of California at San Diego. He received a Ph.D. in Computational Sciences and Informatics in 1996 from George Mason University under the direction of Edward J. Wegman. From 1985-1994 he worked at the Naval Ocean Systems Center in San Diego doing research on pattern recognition and computational statistics. In 1994 he moved to the Naval Surface Warfare Center in Dahlgren, Virginia where he does research in computational statistics and pattern recognition, primarily applied to image processing and automatic target recognition. He is currently involved in the application of statistical pattern recognition techiques to the problem of computer intrusion detection.

Bradley Wallet has a BS in Mathematics from Hampden-Sydney College and a MS in Statistical Sciences from George Mason University. In addition, he is a doctoral candidate Computational Sciences and Informatics at GMU. Mr. Wallet has approximately 20 publications. He is currently Director of Pattern Recognition Research for Chroma, Inc. in Burlingame, California. His research interests include massive data sets, dimensionality reduction, selection bias, visualization of 3-D spatial data, seismic processing and AVO analysis, and automated and semi-automated approaches to networkintrusion detection.

Jaimyoung Kwon, Peter Bickel, and John Rice (poster, Hidden Markov Modeling of Freeways Traffic Status)
Jaimyoung Kwon earned the B.S. degree in Computer Science and Statistics from Seoul National University in 1994 and the M.S. in Statistics from Seoul National University in 1996. He is currently a doctoral candidate in Statistics at UC Berkeley, working under the direction of Prof. Peter Bickel. His research interests include nonparametric estimation, model selection and bootstrap in time series context. Since 1997, he has been also involved in research on traffic flow through a large interdisciplinary research group, Partners in Advanced Highways, or PATH and the National Institute of Statistical Science, working with other scientists on various problems in transportation science, including forecasting travel times on freeways and macroscopic modeling of freeway traffic status using hidden Markov model.