Computational Intelligence: Engineering of Hybrid Systems

Negoita M., Neagu D., Palade V.3-540-21858-0

Hybrid Intelligent Systems has become an important research topic in computer science and a key application field in science and engineering. This book offers a gentle introduction to the engineering aspects of hybrid intelligent systems, also emphasizing the interrelation with the main intelligent technologies such as genetic algorithms evolutionary computation, neural networks, fuzzy systems, evolvable hardware, DNA computing, artificial immune systems. A unitary whole of theory and application, the book provides readers with the fundamentals, background information, and practical methods for building a hybrid intelligent system. It treats a panoply of applications, including many in industry, educational systems, forecasting, financial engineering, and bioinformatics. This volume is useful to newcomers in the field because it quickly familiarizes them with engineering elements of developing hybrid intelligent systems and a wide range of real applications, including non-industrial applications. Researchers, developers and technically oriented managers can use the book for developing both new hybrid intelligent systems approaches and new applications requiring the hybridization of the typical tools and concepts to computational intelligence.

---

Table of contents :
front-matter.pdf……Page 1
1.1 “Computational” and “Artificial” Intelligence Distinguished……Page 21
1.2 Key Features of Intelligent Systems……Page 24
1.3 Classification of Hybrid Intelligent Systems……Page 28
2.1 Introduction……Page 32
2.2 Methods of NN-FS Hybridization in Fault Diagnosis……Page 33
2.2.1 Neuro-Fuzzy Networks……Page 34
2.2.2 Residual Generation Using Neuro-Fuzzy Models……Page 37
2.2.3 Neuro-Fuzzy Based Residual Evaluation……Page 40
2.3 Case Example: Fault Detection and Isolationof an Electro-Pneumatic Valve……Page 41
3 Neuro-Fuzzy Integration in Hybrid Intelligent Systems……Page 44
3.1 Neuro-Fuzzy Integration……Page 46
3.1.1 The Fuzzy Neuronal Model……Page 47
3.2 Knowledge Representationin Hybrid Intelligent Systems……Page 50
3.2.1 The Implicit Knowledge Representation……Page 51
3.2.2 The Explicit Knowledge Representation……Page 54
3.3 Concluding Remarks……Page 58
4.1 Introduction……Page 59
4.2.1 The Notion of f-duality……Page 60
4.2.2 Fuzzy Interactive iatan-OR Operator Used for Rules Extraction from ANNs……Page 61
4.2.3 Fuzzy Interactive itanh-OR Operator Usedfor Rules Extraction from ANNs……Page 65
4.3.1 Interactive Operators for the IRIS Problem……Page 69
4.3.2 Interactive Operators for the Portfolio Problem……Page 72
4.4 Concluding Remarks……Page 75
5.1 Introduction……Page 76
5.2 A Formal Description of Implicitand Explicit Knowledge-Based Intelligent Systems……Page 77
5.3.1 Fire Each Module Method for Integration of Neural Implicit and Explicit Knowledge Modules……Page 79
5.3.2 Unsupervised-Trained Gating Network Methodfor Integration of Neural Implicitand Explicit Knowledge Modules……Page 81
5.3.3 Supervised-Trained Gating Network Methodfor Integration of Neural Implicitand Explicit Knowledge Modules……Page 84
5.4 Concluding Remarks on Explicitand Implicit Knowledge Integration……Page 85
6 Practical Implementation Aspects Regarding Real-World Application of Hybrid Intelligent Systems……Page 87
6.1 An Overview on Application Areaof Hybrid Intelligent Systems……Page 88
6.2 WITNeSS — (Wellington Institute of Technology Novel Expert Student Support) — An Intelligent Tutoring System (ITS)……Page 91
6.2.1 Introduction on ITS……Page 95
6.2.2 Why ITS Are Required……Page 96
6.2.3 The Basic ITS Structure……Page 97
6.2.4 Drawbacks and Problems within the Field of ITS……Page 98
6.2.5 WITNeSS — A Fuzzy Neural Soft Computing Systemfor Optimizing the Presentation of Learning Materialto a Student……Page 100
6.2.6 The Optimiser Agent in WITNeSS……Page 106
6.2.7 Concluding Remarks on WITNeSS……Page 109
6.3 GA Relied Hybrid Intelligent Systemsfor Optimisation Objectives of Fuzzy Systemsand Neural Networks……Page 110
6.3.1 GA with Variable Length Chromosomesas an Optimisation Toolin Complex Hybrid Intelligent Systems……Page 111
6.3.2 VLGGA Based Learning the Parametersof a Fuzzy Inference System from Examples……Page 112
6.3.3 VLGGA for Optimization of Fuzzy Recurrent NN……Page 115
6.4 An Original Neuro-Fuzzy Knowledge-Based Systemfor Air Quality Prediction: NEIKeS (Neural Explicitand Implicit Knowledge Based System)……Page 121
6.4.2 Related Work……Page 122
6.4.3 Data Preparation……Page 123
6.4.4 The Neuro-Fuzzy Knowledge-Based Systemfor Air Quality Predictionfor Air Quality Prediction: NEIKES……Page 130
6.4.5 Results……Page 135
6.4.6 A Distributed AI Architecturefor Air Quality Monitoring……Page 140
6.4.7 Conclusions……Page 143
6.5.1 Introduction……Page 144
6.5.2 Predictive Toxicology……Page 145
6.5.3 Applications of Hybrid Intelligent Systemsin Predictive Toxicology……Page 147
6.5.4 Conclusions……Page 165
7.1 A Brief Introduction to AIS Computing Framework……Page 167
7.2 A Survey on AIS Applicationto Hybrid Intelligent Systems……Page 169
7.3 A Brief Introduction to Fundamentals of DNA Computing……Page 173
7.4 A Survey on DNA Applicationsto Hybrid Intelligent Systems……Page 175
8.1.1 GA as Tools to Design Fuzzy Systems……Page 178
8.1.2 GA Based Hybridization for Knowledge Handling in Expert Systems (Fuzzy Rules Learning and Selection)……Page 179
8.1.3 GA Based HIS for Solving Relational Equations……Page 182
8.1.4 GA Based HIS for Prototype Based Fuzzy Clustering……Page 186
8.1.5 GA Based HIS in Fuzzy Modeling and Systems Identification……Page 187
8.2 HIS Design Methods for High Performance GA……Page 191
8.3.1 GA Based HIS Methods for NN Training……Page 192
8.3.2 HIS for NN Compiling into Fuzzy Rules Using GA……Page 193
8.3.3 Advanced Methods of GA Based HIS Design of NN……Page 197
8.4.1 GA Based Hybridization in Pattern Recognitionand Image Processing……Page 199
8.4.2 GA Based HIS in Different Forecasting Applications……Page 200
8.4.3 GA Based HIS in Job Shop Scheduling……Page 201
8.5.1 EHW Definitionand General Consideration — Implications on Engineering Design and Automation……Page 202
8.5.3 EHW Architectures — The Ideal Environmentfor Implementing the Hardware Hybridization of Different Intelligent Techniques……Page 204
8.5.4 A Suitable GA for Implementing the Hardware Hybridization of Intelligent Techniquesin EHW Architectures……Page 207
8.5.5 GA Based Hybrid Intelligent Systemsand Emergent Technologies. Applicationto NN Based Multi-Classifier Systemsfor Gene Identificationin DNA Sequences……Page 208
back-matter.pdf……Page 210