Muutke küpsiste eelistusi

Intelligent Systems: The Industrial Electronics Handbook [Kõva köide]

Edited by (Auburn University, Alabama, USA), Edited by (Auburn University, Alabama, USA)
  • Formaat: Hardback, 610 pages, kõrgus x laius: 254x178 mm, kaal: 1246 g, 51 Tables, black and white; 357 Illustrations, black and white
  • Ilmumisaeg: 28-Feb-2011
  • Kirjastus: CRC Press Inc
  • ISBN-10: 1439802831
  • ISBN-13: 9781439802830
Teised raamatud teemal:
Intelligent Systems: The Industrial Electronics HandbookSuurem pilt
  • Formaat: Hardback, 610 pages, kõrgus x laius: 254x178 mm, kaal: 1246 g, 51 Tables, black and white; 357 Illustrations, black and white
  • Ilmumisaeg: 28-Feb-2011
  • Kirjastus: CRC Press Inc
  • ISBN-10: 1439802831
  • ISBN-13: 9781439802830
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9781439802830.html
Märksõnad:
The Industrial Electronics Handbook, Second Edition combines traditional and newer, more specialized knowledge that will help industrial electronics engineers develop practical solutions for the design and implementation of high-power applications. Embracing the broad technological scope of the field, this collection explores fundamental areas, including analog and digital circuits, electronics, electromagnetic machines, signal processing, and industrial control and communications systems. It also facilitates the use of intelligent systemssuch as neural networks, fuzzy systems, and evolutionary methodsin terms of a hierarchical structure that makes factory control and supervision more efficient by addressing the needs of all production components.

Enhancing its value, this fully updated collection presents research and global trends as published in the IEEE Transactions on Industrial Electronics Journal, one of the largest and most respected publications in the field.

As intelligent systems continue to replace and sometimes outperform human intelligence in decision-making processes, they have made substantial contributions to the solution of very complex problems. As a result, the field of computational intelligence has branched out in several directions. For instance, artificial neural networks can learn how to classify patterns, such as images or sequences of events, and effectively model complex nonlinear systems. Simple and easy to implement, fuzzy systems can be applied to successful modeling and system control.

Illustrating how these and other tools help engineers model nonlinear system behavior, determine and evaluate system parameters, and ensure overall system control, Intelligent Systems:











Addresses various aspects of neural networks and fuzzy systems Focuses on system optimization, covering new techniques such as evolutionary methods, swarm, and ant colony optimizations Discusses several applications that deal with methods of computational intelligence

Other volumes in the set:











Fundamentals of Industrial Electronics Power Electronics and Motor Drives Control and Mechatronics Industrial Communication Systems
Preface xi
Acknowledgments xiii
Editorial Board xv
Editors xvii
Contributors xxi
PART I Introductions
1 Introduction to Intelligent Systems
1(1)
Ryszard Tadeusiewicz
1.1 Introduction
1(1)
1.2 Historical Perspective
2(1)
1.3 Human Knowledge Inside the Machine---Expert Systems
3(1)
1.4 Various Approaches to Intelligent Systems
4(1)
1.5 Pattern Recognition and Classifications
5(2)
1.6 Fuzzy Sets and Fuzzy Logic
7(1)
1.7 Genetic Algorithms and Evolutionary Computing
8(1)
1.8 Evolutionary Computations and Other Biologically Inspired Methods for Problem Solving
9(1)
1.9 Intelligent Agents
10(1)
1.10 Other AI Systems of the Future: Hybrid Solutions
11
References
11
2 From Backpropagation to Neurocontrol
1(1)
Paul J. Werbos
2.1 Listing of Key Types of Tools Available
1(5)
Backpropagation
Efficient Universal Approximation of Nonlinear Functions
More Powerful and General Decision and Control
Time-Lagged Recurrent Networks
Massively Parallel Chips Like Cellular Neural Network Chips
2.2 Historical Background and Larger Context
6
References
8
3 Neural Network-Based Control
1(1)
Mehmet Onder Efe
3.1 Background of Neurocontrol
1(3)
3.2 Learning Algorithms
4(4)
Error Backpropagation
Second-Order Methods
Other Alternatives
3.3 Architectural Varieties
8(2)
Structure
Neuronal Activation Scheme
3.4 Neural Networks for Identification and Control
10(2)
Generating the Training Data
Determining the Necessary Inputs: Information Sufficiency
Generalization or Memorization
Online or Offline Synthesis
3.5 Neurocontrol Architectures
12(3)
3.6 Application Examples
15(8)
Propulsion Actuation Model for a Brushless DC Motor-Propeller Pair
Neural Network-Aided Control of a Quadrotor-Type UAV
3.7 Concluding Remarks
23
Acknowledgments
23(1)
References
23
4 Fuzzy Logic-Based Control Section
1(1)
Mo-Yuen Chow
4.1 Introduction to Intelligent Control
1(2)
Fuzzy Control
4.2 Brief Description of Fuzzy Logic
3(2)
Crisp Set
Fuzzy Set
4.3 Qualitative (Linguistic) to Quantitative Description
5(1)
4.4 Fuzzy Operations
6(3)
Union
Intersection
Complement
4.5 Fuzzy Rules, Inference
9(3)
Fuzzy Relation/Composition/Conditional Statement
Compositional Rule of Inference
Defuzzification
4.6 Fuzzy Control
12(2)
4.7 Fuzzy Control Design
14(8)
DC Motor Dynamics-Assume a Linear Time-Invariant System
Fuzzy Control
4.8 Conclusion and Future Direction
22
References
22
PART II Neural Networks
5 Understanding Neural Networks
1(1)
Bogdan M. Wilamowski
5.1 Introduction
1(1)
5.2 The Neuron
1(4)
5.3 Should We Use Neurons with Bipolar or Unipolar Activation Functions?
5(1)
5.4 Feedforward Neural Networks
5
References
10
6 Neural Network Architectures
1(1)
Bogdan M. Wilamowski
6.1 Introduction
1(1)
6.2 Special Easy-to-Train Neural Network Architectures
1(12)
Polynomial Networks
Functional Link Networks
Sarajedini and Hecht-Nielsen Network
Feedforward Version of the Counterpropagation Network
Learning Vector Quantization
WTA Architecture
Cascade Correlation Architecture
Radial Basis Function Networks
Implementation of RBF Networks with Sigmoidal Neurons
Networks for Solution of Parity-N Problems
Pulse-Coded Neural Networks
6.3 Comparison of Neural Network Topologies
13(1)
6.4 Recurrent Neural Networks
14
Hopfield Network
Autoassociative Memory
BAM---Bidirectional Autoassociative Memories
References
16
7 Radial-Basis-Function Networks
1(1)
Age J. Eide
Thomas Lindblad
Guy Paillet
7.1 Introduction
1(1)
7.2 Radial-Basis-Function Networks
1(6)
Radial Basis Function
Use of the RBF: An Example
Radial-Basis-Function Network
Learning Paradigms
7.3 Implementation in Hardware
7
Implementation for High-Energy Physics
Implementing a Star Tracker for Satellites
Implementing an RBF network in VHDL
References
11
8 GMDH Neural Networks
1(1)
Marcin Mrugalski
Jozef Korbicz
8.1 Introduction
1(1)
8.2 Fundamentals of GMDH Neural Networks
2(5)
Synthesis of the GMDH Neural Network
Parameters Estimation of the GMDH Neural Network
8.3 Generalizations of the GMDH Algorithm
7(5)
Dynamics in GMDH Neural Networks
Synthesis of the Single-Input Dynamic GMDH Neural Network
Synthesis of the Multi-Output GMDH Neural Network
8.4 Application of GMDH Neural Networks
12(7)
Robust GMDH Model-Based Fault Detection
Robust Fault Detection of the Intelligent Actuator
8.5 Conclusions
19
References
20
9 Optimization of Neural Network Architectures
1(1)
Andrzej Obuchowicz
9.1 Problem Statement
1(3)
9.2 MLP as a Canonical Form Approximator of Nonlinearity
4(1)
9.3 Methods of MLP Architecture Optimization
5(17)
Methods Classification
Bottom-Up Approaches
Top-Down (Pruning) Methods
Discrete Optimization Methods
9.4 Summary
22
References
22
10 Parity-N Problems as a Vehicle to Compare Efficiencies of Neural Network Architectures
1(1)
Bogdan M. Wilamowski
Hao Yu
Kun Tao Chung
10.1 Introduction
1(1)
10.2 MLP Networks with One Hidden Layer
2(2)
10.3 BMLP Networks
4(2)
BMLP Networks with One Hidden Layer
BMLP Networks with Multiple Hidden Layers
10.4 FCC Networks
6(2)
10.5 Comparison of Topologies
8(1)
10.6 Conclusion
8
References
8
11 Neural Networks Learning
1(1)
Bogdan M. Wilamowski
11.1 Introduction
1(1)
11.2 Foundations of Neural Network Learning
1(2)
11.3 Learning Rules for Single Neuron
3(4)
Hebbian Learning Rule
Correlation Learning Rule
Instar Learning Rule
Winner Takes All
Outstar Learning Rule
Widrow-Hoff LMS Learning Rule
Linear Regression
Delta Learning Rule
11.4 Training of Multilayer Networks
7(6)
Error Back-Propagation Learning
Improvements of EBP
Quickprop Algorithm
RPROP-Resilient Error BackPropagation
Back Percolation
Delta-Bar-Delta
11.5 Advanced Learning Algorithms
13(3)
Levenberg-Marquardt Algorithm
Neuron by Neuron
11.6 Warnings about Neural Network Training
16(1)
11.7 Conclusion
16
References
17
12 Levenberg-Marquardt Training
1(1)
Hao Yu
Bogdan M. Wilamowski
12.1 Introduction
1(1)
12.2 Algorithm Derivation
2(6)
Steepest Descent Algorithm
Newton's Method
Gauss-Newton Algorithm
Levenberg-Marquardt Algorithm
12.3 Algorithm Implementation
8(5)
Calculation of the Jacobian Matrix
Training Process Design
12.4 Comparison of Algorithms
13(2)
12.5 Summary
15
References
15
13 NBN Algorithm
1(1)
Bogdan M. Wilamowski
Hao Yu
Nicholas Cotton
13.1 Introduction
1(1)
13.2 Computational Fundamentals
1(4)
Definition of Basic Concepts in Neural Network Training
Jacobian Matrix Computation
13.3 Training Arbitrarily Connected Neural Networks
5(4)
Importance of Training Arbitrarily Connected Neural Networks
Creation of Jacobian Matrix for Arbitrarily Connected Neural Networks
Solve Problems with Arbitrarily Connected Neural Networks
13.4 Forward-Only Computation
9(8)
Derivation
Calculation of δ Matrix for FCC Architectures
Training Arbitrarily Connected Neural Networks
Experimental Results
13.5 Direct Computation of Quasi-Hessian Matrix and Gradient Vector
17(5)
Memory Limitation in the LM Algorithm
Review of Matrix Algebra
Quasi-Hessian Matrix Computation
Gradient Vector Computation
Jacobian Row Computation
Comparison on Memory and Time Consumption
13.6 Conclusion
22
References
23
14 Accelerating the Multilayer Perceptron Learning Algorithms
1(1)
Sabeur Abid
Farhat Fnaiech
Barrie W. Jervis
14.1 Introduction
2(1)
14.2 Review of the Multilayer Perceptron
2(1)
14.3 Review of the Standard Backpropagation Algorithm
3(2)
14.4 Different Approaches for Increasing the Learning Speed
5(2)
Weight Updating Procedure
Principles of Learning
Estimation of Optimal Initial Conditions
Reduction of the Data Size
Estimation of the Optimal NN Structure
Use of Adaptive Parameters
Choice of the Optimization Criterion
Application of More Advanced Algorithms
14.5 Different Approaches to Speed Up the SBP Algorithm
7(4)
Updating the Learning Rate
Updating the Activation Function Slope
14.6 Some Simulation Results
11(3)
Evaluation of the Sensitivity to the Initialization of the Synaptic Weights
Study of the Generalization Capability
Simulation Results and Performance Comparison
14.7 Backpropagation Algorithms with Different Optimization Criteria
14(5)
Modified Backpropagation Algorithm
Least Squares Algorithms for Neural Network Training
14.8 Kalman Filters for MLP Training
19(4)
Multidimensional Kalman Filter Algorithm (FKF)
Extended Kalman Filter Algorithm
14.9 Davidon-Fletcher-Powell Algorithms
23(2)
Davidon-Fletcher-Powell Algorithm for Training MLP
14.10 Some Simulation Results
25(1)
For the 4-b Parity Checker
For the Circle-in-the-Square Problem
14.11 Conclusion
26
Appendix 14.A Different Steps of the FKF Algorithm for Training an MLP
26(2)
Appendix 14.B Different Steps of the EKF for Training an MLP
28(1)
Appendix 14.C Different Steps of the DFP Algorithm for Mathematical Programming
28(1)
Appendix 14.D Different Steps of Wolfe's Line Search Algorithm
29(1)
Appendix 14.E Different Steps of the DFP Algorithm for Training an MLP
30(1)
References
31
15 Feedforward Neural Networks Pruning Algorithms
1(1)
Nader Fnaiech
Farhat Fnaiech
Barrie W. Jervis
15.1 Introduction
1(1)
15.2 Definition of Pruning Algorithms
2(1)
15.3 Review of the Literature
2(1)
15.4 First Method: Iterative-Pruning (IP) Algorithm
3(5)
Some Definitions and Notations
Formulation of the Pruning Problem
How to Choose the Neuron to Be Removed?
15.5 Second Method: Statistical Stepwise Method (SSM) Algorithm
8(1)
Some Definitions and Notations
General Idea
Summary of the Steps in the SSM Algorithm
15.6 Third Method: Combined Statistical Stepwise and Iterative Neural Network Pruning (SSIP) Algorithm
9(3)
First Version: SSIP1
Second Version: SSIP2
15.7 Comments
12(1)
15.8 Simulations and Interpretations
12(1)
15.9 Conclusions
13
Appendix 15.A Algorithm of CGPCNE---Conjugate Gradient Preconditioned Normal Equation
14(1)
References
15
16 Principal Component Analysis
1(1)
Anastasios Tefas
Ioannis Pitas
16.1 Introduction
1(1)
16.2 Principal Component Analysis Algorithm
2(2)
16.3 Computational Complexity and High-Dimensional Data
4(1)
16.4 Singular Value Decomposition
5(1)
16.5 Kernel Principal Component Analysis
6(1)
16.6 PCA Neural Networks
7(1)
16.7 Applications of PCA
7(2)
16.8 Conclusions
9
References
9
17 Adaptive Critic Neural Network Control
1(1)
Gary Yen
17.1 Introduction
1(1)
17.2 Background
2(1)
17.3 Single NN Control Architecture
2(1)
17.4 Adaptive Control Architecture Using Two NNs
3(1)
17.5 Heuristic Dynamic Programming
3(1)
17.6 Dual Heuristic Programming
4(1)
17.7 Globalized Dual Heuristic
5(10)
Introduction
Preliminaries
Identification Neural Network
Action Neural Network
Critic Neural Network
Complete GDHP Algorithm
17.8 Fault Tolerant Control
15(7)
Passive versus Active Approaches
Active FTC Methods
Multiple Model as a Framework
17.9 Case Studies
22(6)
Identification on Using an RNN
FTC Using a GDHP Controller
17.10 Concluding Remarks
28
References
29
18 Self-Organizing Maps
1(1)
Gary Yen
18.1 Introduction
1(5)
Structure
Initialization
Training
Analysis of the Updating Rule
Neighborhood Function
Learning Rate
18.2 Dynamic SOM Models
6(2)
Growing Cell Structure
Growing Neural Gas
Incremental Grid Growing
Other Growing Structure Models
Hierarchical Models
18.3 SOM Visualizations
8(5)
Visualizing Map Topology
Visualizing Data Density
Visualizing Prototype Vectors
Visualizing Component Planes
Visualizing Best Matching Units
Other Visualizations
18.4 SOM-Based Projection
13(8)
User Architecture
Rank Centroid Projection
Selecting the Ranking Parameter R
18.5 Case Studies
21(5)
Encoding of Documents Using Citation Patterns
Collection of Journal Papers on Self-Organizing Maps
Collection of Papers on Anthrax Research
18.6 Conclusion
26
References
27
PART III Fuzzy Systems
19 Fuzzy Logic Controllers
1(1)
Teresa Orlowska-Kowalska
Krzysztof Szabat
19.1 Introduction
1(1)
19.2 Fuzzy versus Classical Control
1(2)
19.3 Fuzzy Models
3(14)
General Structure of Fuzzy Models
Mamdani (Mamdani-Assilian) Model
Takagi-Sugeno Model
Tsukamoto Model
Models with Parametric Consequents
Models Based on Sets of the II-Type Fuzzy Sets
Neuro-Fuzzy Models
Local and Global Models
19.4 Summary
17
References
17
20 Neuro-Fuzzy System
1(1)
Tiantian Xie
Hao Yu
Bogdan M. Wilamowski
20.1 Introduction
1(1)
20.2 Fuzzy System
1(2)
Fuzzification
Fuzzy Rules
Defuzzification
20.3 Neuro-Fuzzy System
3(5)
Structure One
Structure Two
20.4 Conclusion
8
References
8
21 Introduction to Type-2 Fuzzy Logic Controllers
1(1)
Hani Hagras
21.1 Introduction
1(2)
21.2 Type-2 Fuzzy Sets
3(5)
Type-2 Fuzzy Set Terminologies and Operations
21.3 Interval Type-2 FLC
8(2)
Fuzzifier
Rule Base
Fuzzy Inference Engine
Type Reduction
Defuzzilication
21.4 Illustrative Example to Summarize the Operation of the Type-2 FLC
10(4)
Fuzzification
Rule Base
Type Reduction
21.5 Defuzzification
14(1)
21.6 Conclusions and Future Directions
15
References
15
22 Fuzzy Pattern Recognition
1(1)
Witold Pedrycz
22.1 Introduction
1(1)
22.2 Methodology of Fuzzy Sets in Pattern Recognition
2(2)
22.3 Information Granularity and Granular Computing
4(1)
Algorithmic Aspects of Fuzzy Set Technology in Pattern Recognition: Pattern Classifiers
22.4 Fuzzy Linear Classifiers and Fuzzy Nearest Neighbor Classifiers as Representatives of Supervised Fuzzy Classifiers
5(8)
Fuzzy Logic-Oriented Classifiers
Main Categories of Fuzzy Neurons
Architectures of Logic Networks
Granular Constructs of Classifiers
22.5 Unsupervised Learning with Fuzzy Sets
13(6)
Fuzzy C-Means as an Algorithmic Vehicle of Data Reduction through Fuzzy Clusters
Knowledge-Based Clustering
22.6 Data and Dimensionality Reduction
19(1)
22.7 Conclusions
20
Acknowledgments
21(1)
References
21
23 Fuzzy Modeling of Animal Behavior and Biomimcry: The Fuzzy Ant
1(1)
Valeri Rozin
Michael Margaliot
23.1 Introduction
1(2)
Fuzzy Modeling and Biomimicry
23.2 Fuzzy Modeling: A Simple Example
3(2)
23.3 Foraging Behavior of Ants
5(1)
23.4 Fuzzy Modeling of Foraging Behavior
6(2)
Identification of the Variables
Fuzzy Rules
Fuzzy Terms
Fuzzy Inferencing
Parameter Estimation
23.5 Stochastic Model
8(3)
23.6 Averaged Model
11(1)
23.7 Simulations
11(1)
23.8 Analysis of the Averaged Model
12(2)
Equilibrium Solutions
Stability
23.9 Conclusions
14
References
14
PART IV Optimizations
24 Multiobjective Optimization Methods
1(1)
Tak Ming Chan
Kit Sang Tang
Sam Kwong
Kim Fung Man
24.1 Introduction
1(1)
Classical Methodologies
24.2 Multiobjective Evolutionary Algorithms
2(20)
Multiobjective Genetic Algorithm
Niched Pareto Genetic Algorithm 2
Non-Dominated Sorting Genetic Algorithm 2
Strength Pareto Evolutionary Algorithm 2
Pareto Archived Evolution Strategy
Micro Genetic Algorithm Jumping Genes
Particle Swarm Optimization
24.3 Concluding Remarks
22
References
22
25 Fundamentals of Evolutionary Multiobjective Optimization
1(1)
Carlos A. Coello Coello
25.1 Basic Concepts
2(1)
Pareto Optimality
25.2 Use of Evolutionary Algorithms
2(2)
25.3 Multiobjective Evolutionary Algorithms
4(1)
25.4 Applications
5(1)
25.5 Current Challenges
6(1)
25.6 Conclusions
7
Acknowledgment
7(1)
References
7
26 Ant Colony Optimization
1(1)
Christian Blum
Manuel Lopez-Ibdnez
26.1 Introduction
1(1)
26.2 Combinatorial Optimization Problems
2(1)
26.3 Optimization Algorithms
2(1)
26.4 Ant Colony Optimization
3(3)
Solution Construction
Plieromone Update
26.5 Modern ACO Algorithms
6(1)
26.6 Extensions of the ACO Metaheuristic
6(1)
Hybridization with Beam Search
ACO and Constraint Programming
Multilevel Frameworks Based on ACO
26.7 Applications of ACO Algorithms
7(1)
26.8 Concluding Remarks
8
References
8
27 Heuristics for Two-Dimensional Bin-Packing Problems
1(1)
Tak Ming Chan
Filipe Alvelos
Elsa Silva
J.M. Valerio de Carvalho
27.1 Introduction
1(1)
27.2 Bin-Packing Problems
2(3)
27.3 Heuristics
5(9)
One-Phase Heuristics
Two-Phase Heuristics
Local Search Heuristics
27.4 Computational Results
14(3)
27.5 Conclusions
17
References
17
28 Particle Swarm Optimization
1(1)
Adam Slowik
28.1 Introduction
1(1)
28.2 Particle Swarm Optimization Algorithm
2(1)
28.3 Modifications of PSO Algorithm
3(2)
Velocity Clamping
Inertia Weight
Constriction Coefficient
28.4 Example
5(4)
Random Creation of the Population P Consisting M Particles
Evaluation of Particle Positions Using Objective Function FC
Calculation of the Best Neighbors (Only for LPSO Algorithm)
Calculation of New Values of Particle Velocity Calculation of New Values of Particle Position Vectors
28.5 Summary
9
References
9
PART V Applications
29 Evolutionary Computation
1(1)
Adam Slowik
29.1 Introduction
1(1)
29.2 Description of Evolutionary Algorithms
1(8)
Fitness Function
Representation of Individuals---Creation of Population
Evaluation of Individuals
Selection
Mutation and Cross-Over
Terminate Conditions of the Algorithm
Example
29.3 Conclusions
9
References
9
30 Data Mining
1(1)
Milos Manic
30.1 Introduction
1(1)
30.2 What Is Data Mining?
2(1)
30.3 OLAP versus OLTP
3(4)
Data Cubes
OLAP Techniques on Data Cubes
30.4 Data Repositories, Data Mining Tasks, and Data Mining Patterns
7(3)
Data Repositories
Data Mining Tasks
Data Mining Patterns
30.5 Data Mining Techniques
10(2)
Regression Analysis
Decision Trees
Neural Networks
30.6 Multidimensional Database Schemas
12(1)
30.7 Mining Multimedia Data
12(1)
30.8 Accuracy Estimation and Improvement Techniques
13(1)
Accuracy Measures
Accuracy Improvement
30.9 Summary
14
References
14
31 Autonomous Mental Development
1(1)
Juyang Weng
31.1 Biological Development
1(1)
31.2 Why Autonomous Mental Development?
2(2)
31.3 Paradigm of Autonomous Development
4(1)
31.4 Learning Types
5(1)
31.5 Developmental Mental Architectures
6(9)
Top-Down Attention Is Hard
Motor Shapes Cortical Areas
Brain Scale: "Where" and "What" Pathways
System
Pathway Scale: Bottom-Up and Top-Down
Cortex Scale: Feature Layers and Assistant Layers
Level Scale: Dually Optimal CCI LCA
31.6 Summary
15
References
15
32 Synthetic Biometrics for Testing Biometric Systems and User Training
1(1)
Svetlana N. Yanushkevich
Adrian Stoica
Ronald R. Yager
Oleg Boulanov
Vlad P. Shmerko
32.1 Introduction
1(1)
32.2 Synthetic Biometrics
2(2)
Synthetic Fingerprints
Synthetic Iris and Retina Images
Synthetic Signatures
32.3 Example of the Application of Synthetic Biometric Data
4(4)
Hyperspectral Facial Analysis and Synthesis in Decision-Support Assistant
Hyperspectral Analysis-to-Synthesis 3D Face Model
32.4 Synthetic Data for User Training in Biometric Systems
8(2)
32.5 Other Applications
10
References
11
Index 1
Bogdan M. Wilamowski, J. David Irwin