17th International Workshop on Database and Expert Systems Applications (DEXA'06) (2006)
Sept. 4, 2006 to Sept. 8, 2006
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/DEXA.2006.64
Hugh M. Cartwright , Oxford University, UK
Rohan M Gunatillake , Oxford University, UK
Les Sztandera , Philadelphia University, U.S.A
A crucial task in polymer chemistry is the formulation of materials which satisfy strict property constraints. This paper describes the use of dynamic adaptive agents in the design of novel polymers. A neural network is used to predict the properties of a proposed polymer from its composition, while a genetic algorithm solves the inverse problem by acting as a search agent to find promising formulations. The resulting Hybrid Intelligent System provides a computationally efficient means of searching the virtual space defined by the set of all feasible polymers. We report here results from two initial studies.
H. M. Cartwright, L. Sztandera and R. M. Gunatillake, "Genetic Algorithms and Neural Networks in the Molecular Design of Novel Fibers," 17th International Workshop on Database and Expert Systems Applications (DEXA'06)(DEXA), Krakow, Poland, 2006, pp. 598-602.