Ebook: Fuzzy Logic and Neural Networks for Hybrid Intelligent System Design

This book covers recent developments on fuzzy logic, neural networks and optimization algorithms, as well as their hybrid combinations. In addition, the above-mentioned methods are applied to areas such as intelligent control and robotics, pattern recognition, medical diagnosis, time series prediction and optimization of complex problems. Nowadays, the main topic of the book is highly relevant, as most current intelligent systems and devices in use utilize some form of intelligent feature to enhance their performance. In addition, on the theoretical side, new and advanced models and algorithms of type-2 and type-3 fuzzy logic are presented, which are of great interest to researchers working on these areas. Also, new nature-inspired optimization algorithms and innovative neural models are put forward in the manuscript, which are very popular subjects, at this moment. There are contributions on theoretical aspects as well as applications, which make the book very appealing to a wide audience, ranging from researchers to professors and graduate students working in the theory and applications of the computational intelligence area.

Inspired by the flight behavior and mating process of mayflies, the Mayfly algorithm combines the main advantages of swarm intelligence and evolutionary algorithms, resulting in better performance than the particle swarm algorithm. So, we proposed a modification of Mayfly algorithm by applying a fuzzy parameter adapter to be able to apply this to neural network problems. We were able to observe that the fuzzy adapter improves the speed of convergence of the mayfly algorithm and when applied to a neural network for the Mackey glass series, it manages to detect the optimal number of neurons of the hidden layer for the network architecture. However, when using the Mayfly algorithm to optimize the architecture of neural networks, the results do not improve much, so we can deduce that this metaheuristic method is not recommended (for the moment) for this type of optimization, due to the fact that the root mean square error did not get below 0.001 even using the modified Mayfly algorithm with the fuzzy adapter.

There are a total of 14 papers forming the book in the above-mentioned topics. In conclusion, the edited book comprises papers on diverse aspects of fuzzy logic, neural networks and nature-inspired optimization meta-heuristics for designing and implementing hybrid intelligent systems and their application in areas , such as intelligent control and robotics, pattern recognition, time series prediction and optimization of complex problems.