Integration of genetic algorithms and fuzzy logic into a neural network simulation environment

NEUROGRAPH is a simulation environment for neural networks. It provides an easy to use graphical user interface to design, construct and execute neural networks. The most important design goals were easy extensibility, good performance and the ability to solve real world problems. Extensibility is ensured by an extremely flexible, hierarchical internal representation (data structure), which can be easily manipulated by graphical tools. Current available extensions to the neural network simulator are components for genetic algorithms and fuzzy logic. As an example the automated generation of a neural network by a genetic algorithm is shown. This data structure can be interpreted for debugging purposes or can be executed directly for high performance computing. In the later case C or C++ code is generated from the internal representation of the neural functions, compiled and stored in an executable process (library). NEUROGRAPH has a file or process interface to interconnect which other processes. It is also possible to generate C or C++ code representing the neural network which can be incorporated in other applications