A Proposal of Interactive Genetic Network Programming by Fuzzy Structural Modeling

In order to develop a complex software system, complete and consistent design and implementation methods are required. The automatic design methods of complex systems by evolutionary optimization have been developed. Especially, genetic network programming (GNP), which consists of many nodes, is an effective method of network structure oriented evolutionary methods. In this paper, we propose the Interactive GNP method to obtain a network structure which is fit to feed to the human system designer. The proposed method has two parts; the part depicting network structure by fuzzy structural modeling (FSM) method and the part of interactive GNP which can annihilate/add a node by interactive evolutionary search. To verify the effectiveness of the proposed method, we report experimental results.