Regards,
Jeroen Donkers.
--------------------------------
POSITION AT THE DEPARTMENT OF COMPUTER SCIENCE, UNIVERSITEIT
MAASTRICHT, THE NETHERLANDS
RESEARCH ASSISTANTSHIP LEADING TO Ph.D.
The Computer Science Department of the Universiteit Maastricht in the Netherlands
performs research in the field of Artificial Intelligence. The main subjects of
research are Neural Networks and Multi-Agent Systems. The Department also
participates in various projects, such as the INDEKS project. In this project we have
an open slot for a research assistantship.
INFLUENCE DIAGRAMS IN THE ESTUARIES AND COASTS DSS (INDEKS)
The INDEKS-project is a part of the LWI-program (Land Water Environment
Information Technology) that deals with applying information technology to large-
scale infrastructural projects, for instance the development of the Westerschelde
estuary. During the evaluation of plans for the socio-economical and eco-
morphological development of coasts and estuaries, many factors are involved. The
values of these factors depend on facts (measurements), assumptions (based on a
theory), and on experience (heuristics). Although most factors are uncertain factors,
decision-makers are eager to come to a founded choice. Mathematics provides us
with a large set of statistical methods to support this. However, in domains where the
number of uncertain factors becomes too large, standard statistical techniques do not
lead to good results.
Since 1986 important developments took place in Artificial Intelligence in the field
of reasoning and decision-making with uncertainty. Especially, research on Bayesian
belief networks and influence diagrams has led to new results. Bayesian belief
networks are representations of inferences based on conditional probabilities laid
down by the system. Influence diagrams are extensions of Bayesian belief networks
in which decision moments can be included. At this moment Influence diagrams are
operational in decision support systems. However, they have, not yet been applied to
large-scale projects. At our Department, research is performed since 1996 on an
alternative, viz. Markov decision networks. In contrast to influence diagrams, these
networks can be used to model and solve decision problems that allow repeated
decisions.
The candidate will develop and apply influence diagrams within the LWI-domain. The
research focuses on the main question: Can influence diagrams and Markov decision
networks be used effectively to support large-scale projects with many uncertain
factors? Three questions transpire immediately:
- Which of these two techniques is more reliable?
- What are the criteria applied for these techniques?
- What will be the benefit when using these techniques in large-scale projects with
many uncertain factors?
Candidates should have:
- MSc in computer science, artificial intelligence, or related discipline,
- Interest in theoretical and practical research,
- Good programming skills (e.g., Java and/or C, C++),
- Familiarity with knowledge-acquisition techniques.
Submissions should be sent to the administrator of the Department of Computer
Science, Ms. M. van der Mee, by Email: mee@cs.unimaas.nl or by regular mail:
Department of Computer Science, P.O. Box 616, 6200 MD Maastricht, The
Netherlands.
Additional information can be obtained from Professor H. Jaap van den Herik
(herik@cs.unimaas.nl); +31 43 388 3477) or Jeroen Donkers M.Sc.
(donkers@cs.unimaas.nl); +31 43 3883481).
A salary commensurate with Ph.D. student status in the Netherlands is offered: Dfl.
2135 per month in the first year, rising to Dfl. 3812 (gross) per month in the fourth
year.
-----------------------------------------------------------------
drs H.H.L.M. Donkers
Computer Science, FdAW, Universiteit Maastricht
email: donkers@cs.unimaas.nl http://fanth.cs.unimaas.nl/
tel: (+31) 43 - 3883481 fax: (+31) 43 - 3252392