摘要
類神經網路軟體於系統識別與控制之研究
研究生：黃威龍
指導教授：楊世銘
倒傳遞類神經網路是目前學習模式中最具代表性，應用最普遍的類神經網路。但是倒傳遞網路有需要長時間訓練以及容易收斂於局部最小值的缺點。為了改善這些缺點，本文提出擬牛頓學習法和基因演算學習法，並將此結合於類神經網路軟體(VIBNN)中。此研究主旨在發展一套整合性類神經網路視窗軟體，以改善倒傳遞類神經網路的學習效率和精度。本文另提出兩步驟訓練法，其結合擬牛頓法以及基因演算法之優點。本文應用VIBNN於系統識別和振動控制，結果顯示VIBNN可以成功應用於這些問題上。

Development of an Artificial Neural Network Software
for System Identification and Control
Abstract
It is known that backpropagation network requires long training time and is easily stuck in local minimum. To improve these disadvantages, the combination of neural network, quasi-Newton algorithm and genetic algorithm is integrated in the neural network software VIBNN. The purpose of this thesis is mainly to provide an integrated software for improving the training efficiency and accuracy of backpropagation network. A two-stage training algorithm is developed which includes the advantages of quasi-Newton algorithm and of genetic algorithm. Applications of the neural network software to system identification and vibration suppression in this thesis have been very successfully.

Table of Contents
Abstrac … …………………………………………………… i
Table of Contents ……………………………………………… ii
Chapter 1 Introduction
1.1 Motivation …… ………………………………………… 1
1.2 Literature Review …………………………………………… 2
1.2.1 Backpropagation Algorithm ……… …………….……. 2
1.2.2 Neural Network with Genetic Algorithm………………... 3
1.2.3 Neural Network for System Identification and Control .…. 4
1.3 Outline ……………………………………………………. 5
Chapter 2 Artificial Neural Networks
2.1 Introduction ………………………………………………. 6
2.2 Neural Network ………………………………… ………... 7
2.2.1 Feedforward Network with Backpropagation Algorithm … 8
2.2.2 Feedforward Network with Quasi-Newton Algorithm . 11
2.2.3 Feedforward Network with Genetic Algorithm …… 13
Chapter 3 System Identification Using Neural Networks
3.1 Introduction ………… ……………………………… 23
3.2 Introduction to VIBNN ……………… ……………… 25
3.3 Examples of System Identification …………………… 27
3.4 The Two-Stage Training Algorithm …………………… 32
3.5 Conclusion ………………………………………… . 33
Chapter 4 Controller Design by Using Neural Network
4.1 Introduction ……………………………………… .. 52
4.2 Linear Reference Model ……………………………… 52
4.3 VIBNN for Neural Controller………….. …………… 53
4.4 Examples of Controller Design by VIBNN …………… 54
4.5 Comparison of VIBNN with Other Softwares ………… 56
4.6 Conclusion………………………………………… 57
Chapter 5 Summary and Conclusions ……………………… 75
References ………………………………………………… 76

References
Baba, N., 1989, "A New Approach for Finding the Global Minimum of Error Function of Neural Networks," Neural Networks, vol. 2, pp. 367-373.
Billing, S. A. and Zheng, G. L., 1995, "Radial Basis Function Network Configuration Using Genetic Algorithms," Neural Networks, vol. 8, no. 6, pp. 877-890.
Botha, E. C., Barnard, E. and Barnard, C. J., 1996, "Feature-Base Classification of Aerospace Radar Targets Using Neural Networks," Neural Networks, vol. 9, no. 1, pp. 129-142.
Cho, S. B., 1997, "Neural-Network Classifiers for Recognizing Totally Unconstrained Handwritten Numerals," IEEE Trans. Neural Networks, vol. 8, no. 1, pp. 43-53.
Farag, W. A., Quintana, V. H. and Germano, L. T., 1998, "A Genetic-Based Neuro-Fuzzy Approach for Modeling and Control of Dynamical System," IEEE Trans. Neural Networks, vol. 9, no. 5, pp. 756-767.
Funahashi, K. I., 1989, "On the Approximate Relization of Continuous Mappings by Neural Networks," Neural Networks, vol. 2, pp. 183-192.
Hagan, M. T. and Menhaj, M. B., 1994, "Training Feedforward Networks with the Marquardt Algorithm," IEEE Trans. Neural Networks, vol. 5, no. 6, pp. 989-993.
Hecht-Nielsen, R., 1987, "Kolmogorov''s Mapping Neural Network Existence
Theorem," Proc. IEEE ICNN, IEEE Press, New York, vol. 3, pp. 11-13.
Hecht-Nielsen, R., 1989, "Theory of the Backpropagation Neural Network," IJCNN, vol. 1, pp. 593-605.
Hung, S. L. and Adeli, H., 1994, "A Parallel Genetic/Neural Network Learning Algorithm for MIMD Shared Memory Machines," IEEE Trans. Neural Networks, vol. 5, no. 6, pp. 900-909.
Jacobs, R. A., 1988, "Increased Rates of Convergence Through Learning Rate Adaptation," Neural Networks, vol. 1, pp. 295-307.
Karakasoglu, A., Suharsanan, S. I. and Sudareshan, M. K., 1993, "Identification and Decentralized Adaptive Control Using Dynamical Neural Networks with Application to Robotic Manipulators," IEEE Trans. Neural Networks, vol. 4, no. 6, pp. 919-930.
Ku, C. C. and Lee, K. Y., 1995, "Diagonal Recurrent Neural Networks for Dynamic Systems Control," IEEE Trans. Neural Networks, vol. 6, no. 1, pp. 144-155.
Levin, A. U. and Narendra, K. S., 1996, "Control of Nonlinear Dynamical Systems Using Neural Networks-Part Ⅱ: Observability, Identification, and Control," IEEE Trans. Neural Networks, vol. 7, no. 1, pp. 30-41.
Lo, S. C., Chan, H. P., Lin, J. S., Li, H., Freedom, M. T. and Mun, S. K., 1995, "Artificial Convolution Neural Network for Medical Image Pattern Recognition," Neural Networks, vol. 8, no. 7-8, pp. 1201-1214.
Magoulas, G. D., Vrahatis, M. N. and Androulakis, G. S., 1997, "Effective Backpropagation Training with Variable Stepsize," Neural Networks, vol. 10, no. 1, pp. 69-82.
Maniezzo, V., 1994, "Genetic Evolution of the Topolgy and Weight Distribution of Neural Networks," IEEE Trans. Neural Networks, vol. 5, no. 1, pp. 39-53.
Miachalewicz, Z., 1996, Genetic Alogorithms + Data Structure = Evolution Programs, New York: Springer-Verlag.
Minai, A. A. and Williams, R. D., 1990, "Acceleration of Back-propagation Through Learning Rate and Momentum Adaptation," Proc. of IJCNN, vol. 1, pp. 676-679.
Narendra, K. S. and Parthasarathy, K., 1990, "Identification and Control of Dynamical Systems Using Neural Networks," IEEE Trans. Neural Networks, vol. 1, no. 1, pp. 4-27.
Nguyen, D. and Widrow, B., 1990, "Improving the Learning Speed of 2-Layer Neural Networks by Choosing Initial Values of the Adaptive Weights," Proc. of IJCNN, San Diego, California, vol. 3, pp. 22-26.
Parker, D. B., 1985, "Learning Logic," Technical Report TR-47, Center for Computational Research in Economics and Management Science, Massachusetts Institute of Technology, Cambridge, MA.
Phansalkar, V. V. and Sastry, P. S., 1994, "Analysis of the Back-propagation Algorithm with Momentum," IEEE Trans. Neural Networks, vol. 5, no. 3, pp. 505-509.
Roger, S. K., Colombi, J. M., Martin, C. E., Gainey, J. C., Fielding, K. H., Burns, T. J., Ruck, D. W., Kabrisky, M. and Oxley, M., 1995, "Neural Network for Automatic Target Recognition," Neural Networks, vol. 8, no. 7-8, pp. 1153-1184.
Rosenblatt, F., 1958, "The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain," Psych. Rev., vol. 65, pp. 386-408.
Rumelhart, D. E., Hinton, G. E. and Williams, P. J., 1986, "Learning Internal Representation by Error Propagation," in Parallel Distributed Processing, D. E. Rumelhart and J. L. McClelland, Eds. Cambridge, MA: MIT Press.
Schittenkopf, C., Deco, G. and Brauer, W., 1997, "Two strategies to avoid over- fitting in Feed forward Networks," Neural Networks, vol. 10, no. 3, pp. 505-516.
Stager, F. and Agarwal, M., 1997, "Three Methods to Speed Up the Training of Feedforward and Feedback Perceptrons," Neural Networks, vol. 10, no. 8, pp. 1435-1443.
Werbos, P. J., 1974, "Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences," Ph. D. Thesis, Harvard Univ., Cambridge, MA.
Yang, S. M. and Lee, G. S., 1997, "Vibration Control of Smart Structure by Using Neural Networks," Trans ASME. J. of Dynamic Systems, Measurment, and control, vol. 119, no. 1, pp. 34-39.
Yu, X. H., Chen, G. A. and Cheng, S. X., 1995, "Dynamic Learning Rate Optimization of the Backpropagation Algorithm," IEEE Trans. Neural Networks, vol. 6, no. 3, pp. 669-677.
Yu, X. H. and Chen, G. A., 1997, "Efficient Backpropagation Learning Using Optimal Learning Rate and Momentum," Neural Networks, vol. 10, no. 3, pp. 517-527.
Zhu, C. and Paul, F. W., 1995, "A Fourier Series Neural Network and Its Application to System Identification," J. of Dynamic Systems, Measurement, and Control, vol. 117, pp. 253-261.