Object oriented neural networks

Artificial neural networks offer feasible solutions to highly complex problems, such as those encountered in many control problems. Their potency lies primarily in providing a suitable framework for the development of knowledge representation systems. The paper outlines the compatibility and advantages of object-oriented technology for software simulation and application of neural networks. For this, a number of computational models are discussed, particularly in relation to concurrency and distributed processing, real-time performance issues, fault tolerance and accuracy. A brief description of the implementation of the C++ programming language is presented