CO (Complex Systems)

Mantainer

Miguel Rocha

Objectives

Understanding the properties of 'complex systems' with a large number of highly interconnected heterogeneous elements poses today a grand challenge for system research. Examples of such systems include the Internet, critical infrastructures -like computer networks or power grids - and signalling and regulatory networks in biology. For such systems we can design components and their connections but the problem remains of how to guide them to achieve desired global behaviours, like dependability and adaptability, and how to predict and avoid undesired behaviours, like cascading failures in interconnected infrastructures. In the life sciences, novel data acquisition techniques provide a wealth of data on living systems but we lack sufficient means to infer models from these data.

The objective is therefore to develop scalable computational modelling and inference tools and scalable simulation techniques for complex systems and in particular to:

• Infer system models - the dynamic laws governing the components and their interactions - from high volume,

possible incomplete or uncertain data.

• Develop models of emergence of aggregate behaviour that will permit the formulation of design strategies

for systems with a specified aggregate behaviour

Focus/Approach

One or more of the following research issues should be addressed:

• Multi-scale simulations: Develop methods for the effective computation of systems acting/described on different

levels of aggregation.

• Simulation in presence of uncertainty: Develop computational tools that take into account the fact that the

models themselves as well as the parameters that they use may be uncertain.

• Reconstruction of system models from incomplete, missing or inconsistent sets of data.
• Integrated modelling and simulation environments: Matching large amounts of data against models

UM/EEng Competences

Competences:

• Machine Learning / Data Mining
• Bioinformatics
• Modeling and optimization of bioprocesses
• Artificial Neural Networks;
• Evolutionary Computation;
• Networking / Internet

The study of complex systems has not been an area of active research at Univ. Minho but there are basic competences that allow this to be an area to focus in the future. There is a clear need to find partners with competences in the life sciences fields (Biology, Biotechnology, Health care, etc) which are available in our School of Engineering.

-- CarlosBaquero - 27 Sep 2005