Research project P6/04 (Research action P6)
The DYSCO network (IAP VI/4) is a continuation of an existing IAP network (IAP V/22) which has taken as objective to structure and coordinate the research and the graduate training in « systems and control » (in a very broad sense) in Belgium. In the new network, we shall therefore continue to pursue a double objective:
• produce research results that are at the forefront of international research in our areas of excellence within the discipline of “systems and control”
• provide the best possible environment for the training of our PhD students and post-docs.
One of the features of the “systems and control” discipline is that it is an interdisciplinary research area with some specific features and methodologies of its own. The systems approach and the ever-present notion of feedback are two such features. Over the last decade, the progress accomplished by the “systems and control” community in solving many of the core problems of systems and control theory (nonlinear dynamics, modelling, identification, model-predictive control, robust control, optimal control, etc) has led to a spectacular expansion of the tools of our discipline into an ever growing number of other disciplines: environmental systems, communication systems, robotics, networks, graph theory, computer science, biological and bio-chemical systems, biomedical engineering, etc. Each of these fields comes with its own models and hence its own theoretical problems, towards the solution of which systems and control can provide important contributions. In fact, systems and control has become nowadays the single unifying discipline that allows engineers and other scientists to understand the application fields we have mentioned in terms of dynamical systems. However, to achieve significant results in any one of these disciplines requires an in-depth study of their specific models and methods and extensive interactions with the experts of these fields.
The size of DYSCO (about 200 researchers), the presence of several of the leaders of our field including a strong team of internationally renowned experts in applied mathematics, and our specific ways of collaboration allow us to make significant contributions not only in our core discipline of systems and control (including modelling, identification, computation and optimization) but also in several of these neighbouring disciplines in which we have created a critical mass of competent researchers (biological and bio-chemical systems, networks and graphs, traffic systems, biomedical systems, etc). This way of working and of defining our research objectives on a very broad front (core discipline, and several key neighbouring disciplines) has clearly met with success, as observed by the panel of experts who evaluated our network in October 2005. They write: “In terms of numbers of prominent researchers, and in terms of the significance and the volume of its research, the Network remains the pre-eminent research grouping in systems and control, certainly in Europe, and probably in the World”.
Our plan therefore is to maintain the same winning research strategy rather than spelling out in too much detail specific research objectives. This strategy can be summarized as follows. We plan to use the collaborative tools of our network, and the distinguishing presence in our network of top level applied mathematicians (a feature noted by the evaluation panel) to
• pursue fundamental research in the core disciplines of systems and control in which we have leading expertise: optimization, numerical solutions of systems and control problems, modelling and identification, experiment design, networks of nonlinear systems, distributed parameter systems. Compared to the present network, the activity in optimization will be significantly strengthened with the creation of the new Centre of Excellence on Optimization in Engineering at the KULeuven in which our KULeuven teams are involved (see http://www.cs.kuleuven.be/coe-oe/), and the addition of P.A. Absil and F. Glineur in the UCL team. The inclusion of the ULB team (M. Kinnaert) will also allow us to expand our activity in identification towards fault detection and isolation, a new topic for our network that has become increasingly important with the development of large scale fully automated technologies;
• create or develop critical mass of collaborative experts in several disciplines in which systems and control tools can provide significant contributions: chemical and bio-chemical systems, networks and graphs, traffic and transportation problems, motion and coordination problems, biological systems, biomedical engineering. Our already strong activity in bio-chemical and biomedical systems will be strengthened by the inclusion of the new teams at ULB and FPMs (A. Vande Wouwer), which are both heavily involved in the modelling and control of these systems.
We already have critical mass in most of these areas; the addition of the new and emerging teams will reinforce this. In our choice of these research fields, we are comforted by the Murray report (Panel on Future Directions in Control, Dynamics and Systems, see http://www.cds.caltech.edu/~murray/cdspanel/) which identified the major challenges facing systems and control in the future. Our network is actively at work on most of these challenges. In addition, our particular research strategy in which new developments are pushed by the more theoretical researchers in our network and are pulled by those who have ventured into application fields have clearly been endorsed by the evaluation panel of our network. In their 2005 report they write: “By applying system-theoretic approaches in an original way to a broad range of problems (eg synonym detection in dictionaries), and by correctly identifying significant technological developments and their consequences (particularly those due to the continued growth of computing power, and the recent explosive growth of “data in search of theory” from various biological domains), the Network is itself making a major contribution to the continuing timeliness and relevance of systems and control as a scientific discipline. This research domain will therefore remain a key one for the foreseeable future, and certainly for the next 5-10 years.”
On the training side, we plan to continue and strengthen the mechanisms for scientific collaboration and for postgraduate training that we have successfully developed in the previous IAP phases: graduate school in systems and control (which we started in 1991, well before the creation of other graduate schools in Belgium), regular study days, specialized workshops, co-supervision of PhD students by members from different teams, invitation of high level international visitors, focused research groups across teams. The basic goal here is to create a most exciting scientific environment for our PhD students and post-docs. We believe our track record is particularly strong in terms of the quality and breadth of the training of our PhD students and post-docs, and our plan is therefore to maintain the same quality requirements in this area. We are again comforted here by the 2005 evaluation report: “The range and pattern of formal network activities has been very well designed and appears to be extremely effective. The study days and the monthly newsletter ensure that all the personnel in the Network know each other, and are aware of each other’s research topics and achievements. The graduate courses provide an excellent mechanism for ensuring that members of the Network are kept aware of recent developments by leading authorities, both national and international. The Network succeeds in attracting major international figures to Belgium, and being in active collaboration with many of them.”
Footnote on bio-informatics: Bio-informatics research was included and reported upon in our IAP V/22 network "Dynamical Systems and Control: Computation, Identification and Modelling". That phase of the IAP programme has allowed us to create a strong team in bio-informatics, as a subgroup of the former KULeuven/P2 team. Because we consider that the domain has reached a sufficient maturity and that the team has reached sufficient strength, it was decided to file a proposal for a new network specifically devoted to bio-informatics, for which this bio-informatics team is the coordinator. We believe that in doing so we have achieved one of the goals of the IAP programme, which is to favour the emergence of new teams and themes. This implies that the work of the partners on bioinformatics is not included in DYSCO, but in this new IAP network, coordinated by a promoter of our former network.
