The mission of the CI Group in our Department is to perform research on various aspects of computational intelligence and to contribute to the Business Mathematics and Informatics curriculum. Computational Intelligence belongs to the areas of evolutionary computing, fuzzy computing and neurocomputing. As opposed to traditional logic based artificial intelligence techniques, computational intelligence techniques are generally bottom-up, where order and structure emerges from an unstructured beginning.
Below is a list of research projects that the CI Group is working on.
Operational Analysis support to Networked Enabled Capabilities (OANEC)
involved: Paul van Klaveren, Martijn Schut, Gusz Eiben
partners: NLDA Den Helder
‘Information dominance and superiority’, ‘network centric warfare’ and ‘agility and adaptivity’ are terms that have become part of the lexicon associated with the transformation of the military force. The concept of Network Enabled Capabilities (NEC) involves the use of complex, networked systems, consisting of many components that are heterogeneous in functionality and capability with both nonlocal and non-linear interactions and effects. The goal of the research is to investigate the interaction between various network topologies on the one hand and type of agents on the other hand with respect to completing some mission. We like to develop measures of effectiveness or fitness that can be used to evaluate the various configurations.
Adaptive Distributed Intelligent Sensor Systems (ADISS)
involved: Willem van Willigen, Martijn Schut, Gusz Eiben
partners: TNO Den Haag
A trend in sensor systems for surveillance is to go from large single platform static systems to smaller, possibly mobile, systems with more adaptable complementary sensors. Promising is the research being done on fusing measurements from sophisticated dissimilar and distributed sensors while tracking. In the cognitive system model that we use, the hierarchical agents building the most suitable awareness as well as the hierarchically organized sensor management agents work together to adapt the system such that the most utile information is provided at the right
time. The particular topic of research is then how these distributed heterogeneous en hierarchically organized agents work together to achieve this.
MEDEA: New Methods for Detecting Anomalies in Large Databases
involved: Rob Konijn, Vytautas Savickas, Wojtek Kowalczyk
partners: Achmea Zorg
The project aims at developing new techniques for detecting anomalies in large sets records.
We are working on the following issues:
- statistical methods for measuring anomalies in data,
- heuristic search methods for finding anomalies in data,
- efficient implementations of algorithms for finding anomalies big data sets.
We are especially interested in detecting anomalies in databases that are maintained by health insurance companies. Detection of anomalies is important for: reducing fraud, improving factoring processes, increasing the efficiency of processing claims.