MANAGING UNCERTAINTY IN COMPLEX MODELS

 

 

MCSG

MUCM

 

 

Useful Links

The MUCM Community would like to share links to other related disciplines and information.

 

If you have any suggestions for useful links to share with the Community, please contact Jeremy Oakely and these will be made available through this page.

 

 

 

 

SIAM/ASA Journal on Uncertainty Quantification (JUQ)

MASCOT NUM Reasearch Group

Reliability & Uncertainties Group

Uncertainty Quantification in Industrial Analysis & Design Working group

ReDICE

QUEST

 

 

]SIAM/ASA Journal on Uncertainty Quantification (JUQ)

Max Gunzburger, Senior Editor; James Berger, Donald Estep, Editors-in-Chief

 

The SIAM/ASA Journal on Uncertainty Quantification publishes research articles presenting significant mathematical, statistical, algorithmic, and application advances in uncertainty quantification, defined as the interface of complex modeling of processes and data, especially characterizations of the uncertainties inherent in the use of such models. The journal also focuses on related fields such as sensitivity analysis, model validation, model calibration, data assimilation, and code verification. The journal also solicits papers describing new ideas that could lead to significant progress in methodology for uncertainty quantification as well as review articles on particular aspects. The journal is dedicated to nurturing synergistic interactions between the mathematical, statistical, computational, and applications communities involved in uncertainty quantification and related areas.

 

 

MASCOT NUM Research Group

The GDR MASCOT-NUM is a French Research Group (The most significant French Research Group dealing computer experiments) dealing with stochastic methods for the analysis of numerical codes. Its main objective is to coordinate research efforts in this scientific area, which is often called design, modeling and analysis of computer experiments. Its activities involve many techniques of applied mathematics. More specifically, it concerns statistics, probability, computer science, numerical analysis, operational research, mathematical physics, …

 

 

Reliability & Uncertainties Group

The objective of the Reliability & Uncertatinties Group, within the French Statistical Society, is to bring together academic and industrial communities working on the theme of probabilistic and statistical approaches for assessing reliability and control of uncertainties in the physical-industrial and environmental studies.

 

The term "reliability" has long been associated in academic and industrial methods and probabilistic and statistical tools related to the safe operation of industrial components and systems worldwide. The key word "uncertainty" refers to a family of problems where the system behavior is described by a deterministic model (physical, phenomenological, experimental), whose inputs are marred by uncertainty and generally modeled as random variables. In this class of problems, the expected result is a characteristic value of the probability distribution of the output variable of the model, typically a time, a quantile or a probability of exceeding a threshold value.

Although these two specialties are sometimes disjointed in scientific communities, they are linked in the technical practice. From the perspective of the practitioner, reliability studies and studies of quantification and propagation of uncertainties meet the same problem, namely the assessment and management of risks associated with a component or a complex industrial system, environmental, economic . Moreover, very often the result of a study of uncertainty is a probability of exceeding a "critical" value, regarded as a "probability of default".

The more methodological point of view, there are around these two topics and issues common needs, for example relating to the presence of censored or missing data, or the difficulty of estimating quantiles low probability. These often difficult the mathematical point of view problems require significant R & D and create a natural link between researchers and practitioners .

 

 

Uncertainty Quantification in Industrial Analysis & Design Working group

The scientific focus of SIG45 (Special Interest Group) within the ERCOFTAC society (computational fluid dynamics) is on the quantification of output uncertainty in fluid mechanics due to uncertainty of any type of input parameters in a simulation code or in experiments. This includes output variability due to inherent variability in nature (or industrial processes), but also uncertainty due to unknown input parameters (e.g. due to not-yet decided design parameters, or due to unknown model inputs). Techniques to quantify this include polynomial chaos, interval estimation (e.g. based on min-max optimization, or linearization approaches), sensitivity analysis (possibly adjoint based), Monte-Carlo methods, etc. Moreover, depending on the application, many of these techniques may be combined in new and innovative ways.

It is the long term objective of this SIG to coordinate and encourage joint research efforts between the SIG's industrial and academic members (this may include the definition of joint test cases, organization of workshops, etc.) on uncertainty quantification. In this way, we hope to encourage interaction between industry and academic research, stimulate the transfer of academic research to industry, and identify new challenges for research based on the needs encountered in industrial applications.

 

 

ReDICE

The ReDICE Consortium is a research project gathering partners from industry and academia around innovative mathematical methods for the design and analysis of computer experiments. The main research topics include:

  • Metamodel-based optimization, inversion, and related strategies

  • Multi-fidelity metamodels and mixtures of metamodels

  • Computer experiments involving functional data

  • Special kernels and designs

(Free software for kriging is available through this website)

 

 

QUEST

QUEST is a SciDAC Institute that is focused on uncertainty quantification (UQ) in large-scale scientific computations. Our overarching goal is to provide modeling, algorithmic, and general uncertainty quantification (UQ) expertise, together with software tools, to other SciDAC Institutes, SciDAC applications, and Office of Science projects in general—thereby enabling and guiding a broad range of UQ activities in their respective contexts.

QUEST is a collaboration among six institutions with a history of in-depth collaborations on the development, implementation, and use of UQ algorithms/software in challenging high-performance computing (HPC) environments. QUEST members are leading developers of UQ theory, methods, and software in the technical community. They have worked in all aspects of the UQ problem, and have a solid grasp of the challenges and opportunities in this area. They have developed and maintain leading UQ software products that have been applied in HPC environments, with challenging scientific application codes, including climate, geophysics, and combustion.