During the last decade, the resurgence of artificial neural networks, the applicability of fuzzy logic, and the extensive work on linear adaptive filters have had a significant impact on engineering. The adaptive network concept unifies these areas since they all share this worth...they can adapt to changing situations, learn, or approximate, the relationships among the variables of a given system. This simple form of learning is synthesized by the network structure and their adaptation mechanism. The applications of these networks range from simple interpolation, table substitution, modeling, and the synthesis of complex controllers at the cost of almost no design effort. In this work, the adaptability framework is used to present these powerful networks as well as their training or learning mechanisms using clustering, gradient descent, and evolutionary programming.
展开▼
