Workshop Coping with Complexity: Model Reduction and Data Analysis

Lead Research Organisation: University of Leicester
Department Name: Mathematics

Abstract

The problem of model reduction is one of the central problems in multiscale modelling that has recently attracted much attention from the research community across a number of disciplines ranging from applied mathematics to physics, biology and engineering. It has many different aspects and many versions of problem statements and solution methodologies have been created to answer the recent challenges of natural sciences and engineering, from fluid dynamics and chemical engineering to system biology. Coarse-graining theories and computational approaches are becoming a subject of intense research efforts. Such methods provide an effective route for constructing multi-scale models, in other words models which computationally couple phenomena occurring at different scales (macroscopic and microscopic ones). Given that the range of applications of model reduction techniques is huge, and so is the number of methods, special efforts are needed to collect and generalize these results (obtained in different disciplines and formulated in different languages) into a unified and complete toolbox. The main mission of the Workshop is to create a meeting point between researchers from a wide range o disciplines where knowledge about model reduction and coarse-graining for multiscale phenomena will be constructed, obtained and advanced through the dissemination of particular research efforts and achievements.The theme of the workshop is deliberately broad in scope and aims to promote a vigorous exchange of new ideas and fresh methodological perspectives in the increasingly important area of model reduction and coarse graining for multiscale phenomena through a number of dissemination approaches ranging from formal plenary and regular presentations to informal open-ended discussions. The main thematic areas of the workshop in theoretical and computational approaches are:1) Theoretical Approaches (deterministic and stochastic): Invariant manifolds, inertial manifolds, perturbation theory, approximation theory, normal form theory;2) Computational and Algorithmic Approaches: Legacy codes and timesteppers, numerical computation of invariant manifolds, invariant grids, coarse-graining approaches;3) Data analysis and approximation techniques for model reduction: Principal manifolds, dimension reduction methods, data visualization approaches and their applications; 4) Fields of Applications: Non-equilibrium statistical mechanics, kinetic theory, hydrodynamics and mechanics of continuous media, (bio)chemical kinetics, material modelling, bioinformatics, particulate systems, nonlinear dynamics, nonlinear control.Specific areas of study represented in the workshop include dynamical systems, non-equilibrium statistical mechanics, kinetic theory, hydrodynamics and mechanics of continuous media, (bio)chemical kinetics, particulate systems, nonlinear dynamics, nonlinear control, and nonlinear estimation. The interdisciplinary challenge of the workshop theme has attracted many high-level researchers from various countries. Most of the accepted talks are very interdisciplinary, and combine various methods for solution of model reduction problems. Mathematical methods include: dynamical systems approaches (invariant manifolds, inertial manifolds, geometric singular perturbation, perturbative and non-perturbative methods, KAM theory); modern development of PDE (integrable systems, homogenization theory, numerical methods for PDE); methods of approximation theory (machine learning, manifold learning, radial basis functions approach).The Workshop will constitute a step in collecting and advancing the knowledge generated on model reduction and coarse-graining for multiscale phenomena construction through the dissemination of research results and efforts. This new synthesis will be

Planned Impact

The practical aspects of this research thematology are of potential benefit to companies with strong R&D domains such as fine chemicals, pharmaceuticals etc. Model reduction can shorten development time and therefore cut costs and improve overall production policies. Dissemination of these research aspects to companies can be realised through short courses, and through the Process Integration Research Consortium. To achieve these beneficiaries, we need special efforts. Now the organisers of Workshop contact for this purposes Mathematics and Scientific Computing Group, National Physical Laboratory. A representative of this group is a member of A4A6 program committee. The Organising Committee has also contacted the Smith Institute to find more collaborators in Industry. The plan of join actions with the Smith Institute will be prepared. To increase the impact of the Workshop, the Program Committee plans to organise special brainstorming and sandpit session to initiate interdisciplinary and applied projects. Special tutorials and short lecture courses in most important directions of the workshop will be prepared and published. Web page with abstracts, selected presentations and tutorials will be available for open access on the web-server of the University of Leicester. For most important results presented during the Workshop, special press-releases will be prepared and published. Press-releases about selected results and projects will be prepared and delivered. Descriptions of most promising results and projects will be sent to potential industrial partners (direct mail). This Workshop will also benefit researchers across a wide range of disciplines from applied mathematics to natural sciences and engineering. The first group of immediate beneficiaries are those who will participate in the Workshop. The second group consists of researches who will read 1) the works published in the Proceedings Volume and 2) follow-up publications that arisen from the natural synergies that the interactions at the Workshop will create. The third group of academic beneficiaries consists of students (mainly post-graduate) that will either attend the workshop free of charge or will read the relevant publications and will apply some of the works presented/collected to their own research or form broader more interdisciplinary research perspectives. Special sessions for joint grant initiation are included in the workshop program, with brainstorming and special games procedures. These sessions aim to increase the impact of the Workshop on the future research.

Publications

10 25 50
 
Description The goal of this Workshop was to intensify development of constructive methods for complexity reduction in various areas of mathematical modeling, to identify the most promising directions of research, and to promote interdisciplinary and international collaboration in these directions. For these purposes, the Program Committee has assembled a group of people with a wide variety of expertise reflecting the thematically interdisciplinary nature of the workshop, has organized a series of presentations and encouraged fruitful discussions in an informal and casual format.

Main events (the talks of outstanding scientists):

Linking Diffusion Map variables with equation-free multiscale computations
Yannis G. Kevrekidis, Benjamin Sonday and Mikko Haataja
Princeton University, NJ, USA

Aggregation Procedures for Complex Systems
R.S.MacKay
Mathematics Institute, University of Warwick, Coventry

Fast dispersion slow diffusion multi-scale dynamics: theory and computation
Marshall Slemrod
University of Wisconsin, USA

Analysis a Stiff Limit Cycle with CSP: Glycolysis in Saccharomyces cerevisiae
Panayotis D. Kourdis, Ralf Steuer, Dimitris A. Goussis
National Technical University, Athens, Greece
Institute of Theoretical Biology, Humboldt University, Berlin, Germany

Physics-Based Low Order Galerkin Models in Fluid Dynamics & Flow Control
Gilead Tadmor
Northeastern University, Boston, USA
Bernd R. Noack, Michael Schlegel and Oliver Lehmann
Berlin Institute of Technology MB1, Germany
Marek Morzynski
Poznan University of Technology, Poland

Computing realizations of reaction kinetic networks with given properties
G?bor Szederk?enyi, D?vid Csercsik, Katalin M. Hangos
Computer and Automation Research Institute, Hungarian Academy of Sciences

Bregman Divergences for Exploratory Data Analysis
Colin Fyfe
The University of the West of Scotland

Extracting functional dependence from sparse data and dimensionality reduction for potential energy surfaces
Sergei Manzhos, Koichi Yamashita
Department of Chemical Systems Engineering, University of Tokyo, Japan
Tucker Carrington
Department of Chemistry, Queen's University, Kingston, ON, Canada

The proceedings volume is published by Springer in Lecture Notes in Computational Science and Engineering:
Alexander N. Gorban and Dirk Roose (Editors), Coping with Complexity: Model Reduction and Data Analysis, Lecture Notes in Computational Science and Engineering, Springer, Heidelberg-Dordrecht-London-New York, 2010
Exploitation Route This research achievements presented at the Workshop are of potential benefit to companies with strong domains such as fine chemicals, pharmaceuticals etc. Model reduction can shorten development time and therefore cut costs and improve overall production policies. Dissemination of these research aspects to companies can be realised through short courses, and through the Process Integration Research Consortium.
Sectors Aerospace, Defence and Marine,Chemicals,Education,Pharmaceuticals and Medical Biotechnology,Other

URL http://www.springer.com/mathematics/computational+science+%26+engineering/book/978-3-642-14940-5
 
Description The methods disseminated during the Workshop are now used in some application outside academic research: in applied rock physics, in online diagnosis systems for veterinary, in visualisation of genomic data for cancer study.
Sector Chemicals,Digital/Communication/Information Technologies (including Software),Pharmaceuticals and Medical Biotechnology
Impact Types Economic,Policy & public services