Chapter 5 Learning of Type-2 Fuzzy Logic Systems by Simulated Annealing with Adaptive Step Size

摘要

One of the features of fuzzy logic systems is that they can be hybridised with other methods such as neural networks, genetic algorithms and other search and optimisation approaches. These approaches have been proposed to add a learning capability to fuzzy systems to learn from data [6]. Fuzzy systems are good at explaining how they reach a decision but cannot automatically acquire the rules or membership functions to make a decision [10, p.2]. On the other hand, learning methods such as neural networks cannot explain how a decision was reached but have a good learning capability [10, p.2].