Plymouth Engineering Design Centre newsletter

• Sender: owner-newjour@ccat.sas.upenn.edu

Plymouth Engineering Design Centre newsletter

http://www.tech.plym.ac.uk/soc/research/edc/

The Plymouth Engineering Design Centre (PEDC)  newsletter focuses on
Evolutionary and Adaptive Computing in Design and Manufacture as well as
soft computing and computational intelligence technologies. 

The Plymouth Engineering Design Centre (PEDC) is one of seven engineering
design centres that have been established at various academic institutions
within the UK by the Engineering and Physical Science Research Council.
The mission of the Plymouth EDC is to investigate and develop those
adaptive search techniques that are, in the main, analogous to natural
processes;  incorporate co-operative elements and display emergent
behaviour. The objective is to achieve the integration of these techniques
with all stages of the engineering design process, ie from conceptual
through to detailed design. Such techniques typically include genetic
algorithms, evolution strategies, insect colony metaphors, genetic
programming and immune networks.  Strategies involving co-operative
frameworks incorporating combinations of the main elements of these
techniques play a major role in their successful integration with the
conceptual, preliminary / embodiment and detailed design processes.

The Centre adopts a problem oriented approach and maintains a high degree
of collaboration with groups from the the mainstream mechanical, civil,
electrical and electronics industries. The major objectives of the
Centre's research relate to the identification of high-performance
engineering design solutions through global search; the provision of
decision support during the conceptual / preliminary design and global
optimisation during the more deterministic stages of detailed design.
Other complementary soft computing techniques are integrated with the
adaptive Search frameworks to provide information processing capabilities
where required. The main areas of research within the Centre include:

    1.AS techniques for constraint satisfaction and constrained
optimisation 
    2.Genetic programming for data fitting, model calibration and shape
optimisation 
    3.Cluster-oriented genetic algorithm strategies for design problem
decomposition through high-performance region identification 
    4.Qualitative and quantitative evaluation of GA generated design
solutions for multi-criteria satisfaction 
    5.Co-operative systems for mixed discrete / continuous variable
parameter design problems
    6.Multi-level, co-operative systems incorporating preliminary design
models and complex analysis techniques to reduce computational expense
during detailed design optimisation 
    7.Multi-agent systems for information retrieval and design search in
complex design domains 

Real world design problems provide excellent environments to assess the
potential of the various techniques and strategies and to identify
appropriate research direction to ensure successful integration and
application.

e-mail: Ian Parmee, iparmee@plymouth.ac.uk