[SAMSI logo] 19 T.W. Alexander Drive
P.O. Box 14006
Research Triangle Park, NC 27709-4006
Tel: 919.685.9350
Fax: 919.685.9360
[email protected]
 
OpportunitiesProgramsWorkshopsCalendarAffiliatesReports & PublicationsDirectoryAbout
 

2007-08 Program on Risk Analysis, Extreme Events and Decision Theory

Research Foci
Description of Activities

Working Groups:
Adversarial Risk
Bayesian Methods for Extremes
Environmental Risk Analysis
Services Sector Risk
Multivariate Extremes - Applications
Multivariate Extremes - Methodology

Suggested Reading Materials
Further Information

Introduction

This full-year SAMSI program will address fundamental issues in risk analysis and the linked problems associated with extreme events and decision theory. The program will engage researchers from the statistical sciences, applied mathematical sciences including actuarial science, and the decision sciences, including operations research. The goal is to produce genuine impact on the practice of risk analysis and assessment as well as on theory and methodology for extreme events and decision theory.

Background

Over the past several years, there has been a wealth of scientific progress on risk analysis. As the set of underlying problems has become increasingly diverse, drawing from areas ranging from national defense and homeland security to genetically modified organisms to animal disease epidemics and public health to critical infrastructure, much research has become narrowly focused on a single area. It has also become clear, however, that the need is urgent and compelling for research on risk analysis, extreme events (such as major hurricanes) and decision theory in a broader context. Availability of past information, expert opinion, complex system models, and financial or other cost implications as well as the space of possible decisions may be used to characterize the risks in different settings. Integration of expertise developed by researchers in different scientific communities on each of these facets is the objective of this SAMSI program.

Risk analysis and extreme events also carry a significant public policy component, which is driven in part by the increasing stakes and the multiplicity of stakeholders. In particular, policy concerns direct attention not only to the dramatic risks for huge numbers of people associated, for example, with events of the magnitude of Hurricane Katrina or bioterrorism, but also to "small-scale" risks such as drug interactions driven by rare combinations of genetic factors.

On October 27-29, 2005, the National Institute of Statistical Sciences and Iowa State University co-sponsored a Workshop on Overarching Issues in Risk Analysis, held in Ames, IA. Details of the program are available on the NISS web site. The workshop was meant to be a "stock-taking" of exciting new research on risk analysis over the past several years, seeking answers to such questions as: What are the high-leverage gaps? What issues span multiple problem contexts? What kinds of collaborations among researchers in the statistical, applied mathematical and decision sciences and domain scientists are needed to carry out the research?

All of these threads lead to the SAMSI risk analysis program.

Program Leaders: Dipak Dey (Univ. of Connecticut), Stephen Pollock (Michigan), David Rios Insua (Universidad Rey Juan Carlos), Lawrence Brown (Univ. of Pennsylvania, National Advisory Committee Liaison), Richard Smith (UNC-CH, Local Scientific Coordinator), Nell Sedransk (NISS, SAMSI Directorate Liaison)

Science Advisors: David Banks (Duke), Vickie Bier (Univ. of Wisconsin), James Broffitt (Univ. of Iowa), Alicia Carriquiry (Iowa State), Robert Clemen (Duke), Susan Ellenberg (Univ. of Pennsylvania), Herbert Hethcote (Univ. of Iowa), Wolfgang Kliemann (Iowa State), Robert Winkler (Duke), Stan Young (NISS)

Visitors Speakers (confirmed): Tim Bedford (Strathclyde Business School), Lawrence Brown (Univ. of Pennsylvania), Richard Davis (Colorado State University), Dipak Dey (Univ. of Connecticut), Christl Donnelly (Imperial College), Dougal Goodman (The Foundation for Science and Technology, UK), Tom Knutsen (Princeton University), Howard Kunreuther (Univ. of Pennsylvania), Sidney Resnick (Cornell University), David Rios Insua (Universidad Rey Juan Carlos), Richard Smith (Univ. of North Carolina), Zhengjun Zhang (Univ. of Wisconsin)

Confirmed Speakers and Invited Discussants for the Opening Workshop:

Tim Bedford (Strathclyde Business Schoo)
Vicki M. Bier ((University of Wisconsin-Madison)
Richard Davis (Colorado State University)
Dipak Dey (Univ. of Connecticut)
Paul Garthwaite (Open University)
Dougal Goodman (The Foundation for Science and Technology, UK)
James Hammitt (Harvard University)
Jonathan Hosking (IBM)
Tailen Hsing (Ohio State University)
Tom Knutsen (Princeton University)
N. D. Shyamal Kumar (University of Iowa)
Elisabeth Paté-Cornell (Stanford University)
David Rios Insua (Universidad Rey Juan Carlos) Fabrizio Rugierri (CNR-IMATI)
Lianne Sheppard (University of Washington)
Anne Smith (CRAI)
Richard Smith (University of North Carolina)
Thomas Wallsten (University of Maryland at College Park)
Zhengjun Zhang (University of Wisconsin)

Opening Speakers:

Howard Kunreuther (University of Pennsylvania)
David Rios Insua (Universidad Rey Juan Carlos)

Distinguished Lecturer:

Christl Donnelly (Imperial College)

Closing Lecturer:

Ralph Keeney (Fuqua School, Duke University)

Research Foci

Interdisciplinary working groups are relatively certain to be formed around both kinds of events and critical research tasks in theory and methodology, following the already identified interests and the existing momentum. Other working groups may emerge, and other issues with more specialized focus may serve as the basis for small, intensive mid-program workshops.

Critical research tasks for this program include:

Extreme Values: Theory for Multidimensional Extremes. Probability theory and statistical methodology for one-dimensional distributions of extreme values have been developed over the past half-century, but these do not extend easily to higher dimensions. The contexts of extreme events such as natural disasters or man-created events of great magnitude and rarity are both complex and multi-factor in their origin and also high-dimensional in their consequences. Thus the development of a usable probability theory and accompanying statistical methodology is crucial if statistical thinking is to be applied to these kinds of events; a-stable processes form one focus for theoretical research.

Financial Risk: Risk Assessment and Risk Management for Critical Resources, Infrastructure and Energy Markets. Prediction of financial consequences of extreme events is formulated differently in actuarial science, in operations research, in financial mathematics and in risk management and decision sciences. In the example of pricing structures, the mathematical and statistical models and the embedding of data are different in the insurance (actuarial), the industrial (operations research) and investment (financial mathematics) industries. In particular, the incorporation of data in the form of "estimated" costs may be mathematically formalized or not. To a significant extent, costs are not treated as random; and there is no role for uncertainty associate with the costs themselves or with the process of their estimation. Like other applications where second moments are neglected, the consequences for decision-making can be severe. Serious attention to inclusion of uncertainties in the modeling may offer significant improvement in risk analyses.

Experts and Decisions: Prior Elicitation and Modeling with Expert Opinion. Elicitation is the process of formulating a person's knowledge and beliefs about one or more unknowns (parameters) into a (joint) probability distribution for those unknowns in a decision-making setting. Such elicitation may come from one expert or multiple experts. In statistical analysis, elicitation usually arises as a method for specifying the prior distribution for one or more unknown parameters of a statistical model. In the context a prior distribution for a Bayesian analysis, elicitation is the representation of the expert's prior knowledge in a probabilistic form. An elicitation is done well if the distribution that is derived accurately represents the expert's knowledge, regardless of how good that knowledge is. However, to achieve accurate elicitation is by no means straightforward, even for just a single event or hypothesis. In the case of multiple experts, we often adopt the so called "supra Bayesian" approach as our mechanism for combining expert opinion.

Adversarial Risk: Formalizing Analysis of Risk from Intelligent Opponents. Traditional methodology is based on the canonical risk equation: (Likelihood of Attack) * (Consequence) * (1-System Effectiveness) = Risk. However, new business and government scenarios require considering risk analysis that takes into account opponents' intelligence, possible willingness to cooperate and transfer of risk following decisions and actions. Existing decision theoretic abstractions and tools may be inadequate. However, it may be possible to utilize game theoretic modeling of opponents' strategy coupled with negotiation analysis to estimate the relative risks of various strategies. Models may include multiple points of decision and also incorporate risk aversion as a function of the intensity of the threat.

Contextual settings for assessing risk and making decisions include:

Environmental Risk Analysis: Ecological Risk Assessment and Risks associated with Extreme Climatic Events. For ecological risk, data from plants and animals can measure concentrations of biohazard and bioavailable contaminants in ambient media, to evaluate food chain transfers and, ultimately, to predict ecological risk. In this context many risk models have a critical weakness arising from large unobserved variability. For environmental risk, this problem is well-recognized, but still is not solved. The occurrence of extreme environmental disruptions depends jointly on the highly unstable occurrence of these events and the locally, at least, highly variable consequences.

Industrial Risk: Pharmaceutical and Health Risk. In the pharmaceutical industry, critical decision-making occurs early in the drug-development process as a compound is moved from laboratory into animal studies, from pre-clinical to early clinical, continuing onward through to the final submission for FDA approval. Risk continues to be incurred post-approval and post-marketing through the occurrence and reporting of adverse events [putatively] ascribed to drug usage. Risk analysis and post-market risk management depends on data and analysis especially for side effects and drug toxicities. Contributing statistical issues include inference from non-randomized clinical trials, multiplicity and asymmetries between null and alternative hypotheses, analysis of rare events, competing risks.

Description of Activities

Workshops: The Kickoff Workshop will be September 16, 2007 - September 19, 2007. Its principal goal will be to engage a broadly representative segment of the statistical, applied mathematical and decision analyis/operations research communities in formulation and pursuit of specific research activities to be undertaken by the Program Working Groups, discussed above.

Two additional workshops are scheduled for October and for January. The October workshop will focus on the policies and practical decisions made on the basis for risk modeling and risk assessment. The January workshop will consider mathematical theory for extreme values and the consequences of applying these technical approaches in implementing risk models for extreme climate events. In May, the culmination of the research year will be presented in a SAMSI workshop contiguous with the Interface 2008 meeting.

Further Information

Additional information about the program and opportunities to participate in it is available:


 
 

Entire site © 2001-2010, Statistical and Applied Mathematical Sciences Institute. All Rights Reserved.