A general neural-network-based connectionist model, called Fuzzy Neural Network (FNN), is proposed in this book for the realization of a fuzzy logic control and decision system. The FNN is a feedforward multi-layered network which integrates the basic elements and functions of a traditional fuzzy logic controller into a connectionist structure which has distributed learning abilities.In order to set up this proposed FNN, the author recommends two complementary structure/parameter learning algorithms: a two-phase hybrid learning algorithm and an on-line supervised structure/parameter learning algorithm.Both of these learning algorithms require exact supervised training data for learning. In some real-time applications, exact training data may be expensive or even impossible to get. To solve this reinforcement learning problem for real-world applications, a Reinforcement Fuzzy Neural Network (RFNN) is further proposed. Computer simulation examples are presented to illustrate the performance and applicability of the proposed FNN, RFNN and their associated learning algorithms for various applications.

Neural Fuzzy Systems provides a comprehensive, up-to-date introduction to the basic theories of fuzzy systems and neural networks, as well as an exploration of how these two fields can be integrated to create Neural-Fuzzy Systems. It includes Matlab software, with a Neural Network Toolkit, and a Fuzzy System Toolkit.

A general neural-network-based connectionist model, called Fuzzy Neural Network (FNN), is proposed in this book for the realization of a fuzzy logic control and decision system. The FNN is a feedforward multi-layered network which integrates the basic elements and functions of a traditional fuzzy logic controller into a connectionist structure which has distributed learning abilities.In order to set up this proposed FNN, the author recommends two complementary structure/parameter learning algorithms: a two-phase hybrid learning algorithm and an on-line supervised structure/parameter learning algorithm.Both of these learning algorithms require exact supervised training data for learning. In some real-time applications, exact training data may be expensive or even impossible to get. To solve this reinforcement learning problem for real-world applications, a Reinforcement Fuzzy Neural Network (RFNN) is further proposed. Computer simulation examples are presented to illustrate the performance and applicability of the proposed FNN, RFNN and their associated learning algorithms for various applications.

Presenting current trends in the development and applications of intelligent systems in engineering, this monograph focuses on recent research results in system identification and control. The recurrent neurofuzzy and the fuzzy cognitive network (FCN) models are presented. Both models are suitable for partially-known or unknown complex time-varying systems. Neurofuzzy Adaptive Control contains rigorous proofs of its statements which result in concrete conclusions for the selection of the design parameters of the algorithms presented. The neurofuzzy model combines concepts from fuzzy systems and recurrent high-order neural networks to produce powerful system approximations that are used for adaptive control. The FCN model stems from fuzzy cognitive maps and uses the notion of “concepts” and their causal relationships to capture the behavior of complex systems. The book shows how, with the benefit of proper training algorithms, these models are potent system emulators suitable for use in engineering systems. All chapters are supported by illustrative simulation experiments, while separate chapters are devoted to the potential industrial applications of each model including projects in: • contemporary power generation; • process control and • conventional benchmarking problems. Researchers and graduate students working in adaptive estimation and intelligent control will find Neurofuzzy Adaptive Control of interest both for the currency of its models and because it demonstrates their relevance for real systems. The monograph also shows industrial engineers how to test intelligent adaptive control easily using proven theoretical results.

The advent of the computer age has set in motion a profound shift in our perception of science -its structure, its aims and its evolution. Traditionally, the principal domains of science were, and are, considered to be mathe matics, physics, chemistry, biology, astronomy and related disciplines. But today, and to an increasing extent, scientific progress is being driven by a quest for machine intelligence - for systems which possess a high MIQ (Machine IQ) and can perform a wide variety of physical and mental tasks with minimal human intervention. The role model for intelligent systems is the human mind. The influ ence of the human mind as a role model is clearly visible in the methodolo gies which have emerged, mainly during the past two decades, for the con ception, design and utilization of intelligent systems. At the center of these methodologies are fuzzy logic (FL); neurocomputing (NC); evolutionary computing (EC); probabilistic computing (PC); chaotic computing (CC); and machine learning (ML). Collectively, these methodologies constitute what is called soft computing (SC). In this perspective, soft computing is basically a coalition of methodologies which collectively provide a body of concepts and techniques for automation of reasoning and decision-making in an environment of imprecision, uncertainty and partial truth.

Flexible Neuro-Fuzzy Systems is the first professional literature about the new class of powerful, flexible fuzzy systems. The author incorporates various flexibility parameters to the construction of neuro-fuzzy systems. This approach dramatically improves their performance, allowing the systems to perfectly represent the pattern encoded in data. Flexible Neuro-Fuzzy Systems is the only book that proposes a flexible approach to fuzzy modeling and fills the gap in existing literature. This book introduces new fuzzy systems which outperform previous approaches to system modeling and classification, and has the following features: -Provides a framework for unification, construction and development of neuro-fuzzy systems; -Presents complete algorithms in a systematic and structured fashion, facilitating understanding and implementation, -Covers not only advanced topics but also fundamentals of fuzzy sets, -Includes problems and exercises following each chapter, -Illustrates the results on a wide variety of simulations, -Provides tools for possible applications in business and economics, medicine and bioengineering, automatic control, robotics and civil engineering.

This book is concerned with Intelligent Control methods and applications. The field of intelligent control has been expanded very much during the recent years and a solid body of theoretical and practical results are now available. These results have been obtained through the synergetic fusion of concepts and techniques from a variety of fields such as automatic control, systems science, computer science, neurophysiology and operational research. Intelligent control systems have to perform anthropomorphic tasks fully autonomously or interactively with the human under known or unknown and uncertain environmental conditions. Therefore the basic components of any intelligent control system include cognition, perception, learning, sensing, planning, numeric and symbolic processing, fault detection/repair, reaction, and control action. These components must be linked in a systematic, synergetic and efficient way. Predecessors of intelligent control are adaptive control, self-organizing control, and learning control which are well documented in the literature. Typical application examples of intelligent controls are intelligent robotic systems, intelligent manufacturing systems, intelligent medical systems, and intelligent space teleoperators. Intelligent controllers must employ both quantitative and qualitative information and must be able to cope with severe temporal and spatial variations, in addition to the fundamental task of achieving the desired transient and steady-state performance. Of course the level of intelligence required in each particular application is a matter of discussion between the designers and users. The current literature on intelligent control is increasing, but the information is still available in a sparse and disorganized way.