Engineering design and decision support via rule based genetic optimization

This research activity is aimed at generating a computational environment which is capable of supporting the design process at a conceptual level or the enterprise decision support process in real or near real time. The ultimate goal is to provide design and/or decision support which exceeds the capability of a domain expert. The environment is built upon the foundation provided by an evolutionary or genetic search algorithm. Domain specific knowledge is introduced through a rule base which is manipulated directly by the genetic algorithm. The rule base is constructed so that it does not destroy the global nature of the search. The implementation of the rule base is anticipated to both enhance the character of the solution as well as accelerate the solution process. An intelligent genetic design environment will be developed into a prototype system which will be tested on a diverse set of examples. These examples will include structural design and traditional problems such as a maze solution and traveling salesman.