The thesis focuses on in two vital and interrelated aspects of modelling design support systems, they are: how innovative solutions may arise, and the knowledge-base's extension and maintenance. The dilemma 'reproduction versus creativity' is identified as one of the main deadlocks that the design methods' debate, research in Computer Aided Architectural Design, CAAD, and Artificial Intelligence, AI, have faced in the last thirty years. A hybrid approach is then proposed as a means of overcoming these difficulties, where a rudimentary evolving design support environment is developed. It draws inspiration from three areas of Artificial Intelligence: knowledge-base systems, connectionist models, and case-based reasoning (CBR). However, it differs fundamentally from conventional knowledge-base systems, connectionist models and CBR tools, in its architecture, although strongly inspired by these underlying theories. The main benefits and contributions of this hybrid system are an incremental selfextending feature able to minimise substantially the dependency on knowledge engineer intervention, and an interactive support to innovation by augmenting the designer's creativity.