ATM (Asynchronous transfer mode) 非同步傳輸通訊網路已被公認為多媒體傳輸中最佳的選擇，由於它能夠支援不同性質的流量及不同程度的服務品質需求，因此已逐漸取代傳統的傳輸方式。ATM為了達到品質保證必須有一些流量控制的功能，例如呼叫允許控制及壅塞控制等等。由於傳統方法無法有效地估算出需求頻寬及細胞遺失率，為了改善這些缺點，我們加入了人工智慧的技巧去做流量控制及細胞遺失率的預估，使得ATM控制器更有效率。本論文共分成兩大主題：第一個主題是以類神經網路為基礎的呼叫允許控制器之設計及系統效能分析，第二個主題是著重於智慧型網路壅塞控制器之製作。
在第一主題中，我們著重於RBF網路效能之改善及以類神經網路為基礎的呼叫允許控制器之設計。RBF網路雖然廣泛地被應用於建構輸入及輸出的相關函數，但傳統的RBF網路卻遭遇了兩項困難：第一，RBF網路的初始值需要以錯誤嘗試法來決定；第二，當輸出期望值在某一區間有突然性的變化或者是固定值時，傳統的RBF網路效能會比較差。我們提出了一些方法來克服上述的缺點。首先使用新的隱藏層函數，並且採用ART的演算法來決定隱藏層神經元的個數，因此可以得到較佳的初始參數值及加權值，再經由梯度法則去修正這些加權值使得類神經網路的效能
更好。接著我們把ART-RBF網路應用到ATM的呼叫允許控制器上。由於傳統的方法無法準確地估測出所需要的頻寬，往往造成頻寬過剩或細胞遺失率過高的現象。我們利用ART-RBF網路去估測所需要的頻寬，再來決定接受或拒絕新連線的要求。由於我們的方法所估測出的值較為準確，因此更能有效地控制ATM網路上的品質。
在第二主題中，我們研究網路壅塞的方法並且提出了一個結合模糊類神經網路與速率回饋方式的壅塞控制系統。傳統的方法是直接觀測細胞停留在ATM佇列中長度的大小，假如長度大於預設值時，ATM會通知來源端以固定值來調降傳送的速率，然而這個方法很難根據網路壅塞程度而求出一個合適的來源端速率。為了改善此缺點，我們使用模糊類神經網路來控制來源端所送出的速率，經由訓練後，網路可以根據佇列中的長度及其隨時間的變化來預估出細胞遺失比率，進而推算出來源端下一次所應該傳送的速率。
總之，本論文對類神經網路的架構及效能提出了一些改善的方法，並且將類神經網路應用在ATM網路的交通流量控制上。我們提出一些實用的控制機制，並且經由實驗模擬，顯示這些機制比傳統方法有效。

ATM (Asynchronous Transfer Mode) networks were deemed the best choice for multimedia communication. The traditional mode was replaced because ATM can provide varied traffic types and QoS (quality of service). Maintaining QoS, however, requires a flexible traffic control, including call admission control and congestion control. Traditional approaches fail to estimate the required bandwidth and cell loss rate precisely. To alleviate these problems, we employ AI methods to improve the capability of estimated bandwidth and predicted cell loss rate. This thesis aims to apply neural network techniques to ATM traffic control and consists of two parts. The first part concerns a neural-based call admission control, while the second part presents an intelligent congestion control for ATM networks.
In the first part, we focus on the improvement of RBF (Radial basis function) networks and the design of a neural-based call admission control. RBF networks have been widely used for modeling a function from given input-output patterns. However, two difficulties are encountered with traditional RBF networks. One is that the initial configuration of a RBF network needs to be determined by a trial-and-error method. The other is that the performance suffers from some difficulties when the desired output has abrupt changes or constant values in certain intervals. We propose a novel approach to overcome these difficulties. New kernel functions are used for hidden nodes, and the number of nodes is determined automatically by an ART-like algorithm. Parameters and weights are initialized appropriately, and then tuned and adjusted by the gradient descent method to improve the performance of the network. Then, we employ ART-RBF networks to design and implement a call admission control. Traditional approaches fail to estimate appropriately the required bandwidth, leading to a waste of bandwidth or a high cell loss rate. To alleviate the problem, we employ ART-RBF networks to estimate the required bandwidth, and thus a new connection request can then be accepted or rejected. Because of the more accurate estimation on the required bandwidth, the proposed method can provide a better control on quality of service for ATM networks.
In the second part, we propose a neural-fuzzy rate-based feedback congestion control for ATM networks. Traditional methods perform congestion control by monitoring the queue length. The source rate is decreased by a fixed rate when the queue length is greater than a predefined threshold. However, it is difficult to get a suitable rate according to the degree of traffic congestion. We employ a neural-fuzzy mechanism to control the source rate. Through learning, cell loss can be predicted from the current value and the derivative of the queue length. Then an explicit rate is calculated and the source rate is controlled appropriately.
In summary, we have proposed improvements on architecture and performance of neural networks, and applied neural networks to traffic control for ATM networks. We have developed some control mechanisms which, through simulations, have been shown to be more effective than traditional methods.

Contents
Chapter 1 Introduction…1
1.1 Motivations…2
1.2 Organization of this dissertation…3
Chapter 2 ART-RBF Networks…7
2.1 Introduction…8
2.2 ART-RBF Networks…10
2.2.1 Traditional RBF Networks…11
2.2.2 SRBF Networks…12
2.3 Architecture of Proposed Networks…13
2.4 Initial Configuration…15
2.5 Derivation of Initial Parameters…20
2.6 Learning Rules…22
2.7 Simulation Results…25
2.8 Summary…41
Chapter 3 Call Admission Control…44
3.1 Introduction…44
3.2 Model-Based Aggregate Characteristics…45
3.3 Measurement-Based Control…47
3.4 Ratio-Based Control…49
3.5 Weight-Based Control…51
3.6 Neural-Based Control…56
3.7 Performance Comparison…58
3.7.1 Static Case…58
3.7.2 Dynamic Case…60
3.8 Resource Management…67
3.9 Simulation Results…70
3.10 Summary…81
Chapter 4 Congestion Control…82
4.1 Introduction…83
4.2 Traditional Contol Schemes…84
4.2.1 Proportional Rate Control Algorithm…84
4.2.2 Enhanced Proportional Rate Control Algorithm…87
4.3 Overview of the Proposed Controller…88
4.4 A Predictor Based On Neural-Fuzzy Networks…91
4.5 Source Rate Calculation…97
4.6 Simulation Results…100
4.7 Summary…107
Chapter 5 Leaky Bucket Regulator…108
5.1 Introduction…108
5.2 System Overview…110
5.3 Cell Loss Calculation…111
5.4 Learning Rules…114
5.5 Rate Regulation…116
5.6 Simulation Results…117
5.7 Summary…122
Chapter 6 Conclusions…123
6.1 Review of Main Contributions…123
6.2 Future Work…124
Bibliography…126

[1] A. Atai, “A rate-based feedback traffic control in B-ISDN,”
Proceedings of IEEE International Conference on Communications, pp. 1605-1615, May 1994.
[2] The ATM Forum Technical Committee, “Traffic management specification version 4.0,”1996.
[3] J. Aweya, “Neurocontroller for buffer overload control in a packet switch,” IEE Proceedings - Communications, vol. 145, no. 4, pp. 227-223, August 1998.
[4] E. B. Bartlett, D. V. Sridhar and R. C. Seagrave, “An information theoretic approach for combining neural network process models,” Neural Networks, vol. 12, no. 6, pp. 915-926, July 1999.
[5] L. Benmohamed and S. M. Meerkov, “Feedback control of congestion in packet
switching networks: the case of a single congested node ,” IEEE/ACM Transactions on Networking, vol. 1, no. 6, December 1993.
[6] M. Bichsel and P. Seitz, “Minimum class entropy: A maximum information
approach to layered networks,” Neural Networks, vol. 2, no. 2, pp. 133-141, May1989.
[7] S. A. Billings and X. Hong, “Dual-orthogonal radial basis function networks for nonlinear time series prediction,” Neural Networks, vol. 11, no. 3, pp. 479-493, April 1998.
[8] A. W. Branhart, “Baseline Performance Using PRCA Rate-Control,” ATM Forum, 1994.
[9] G. Bugmann, “Normalized Gaussian radial basis function networks,” Neurocomputing, vol. 20, no. 2, pp. 97-110, April 1998.
[10] G. A. Carpenter and S. Grossberg, “ART2: Self-organization of stable category recognition codes for analog input patterns,” Applied Optics, vol. 26, no. 23, pp. 4919-4930, June 1987.
[11] G. A. Carpenter and S. Grossberg, “The ART of adaptive pattern recognition
by a self-organizing neural network,” Computer, vol. 21, no. 3, pp. 77-88, March 1988.
[12] G. A. Carpenter and S. Grossberg, “Fuzzy ART: Fast stable learning and categorization of analog patterns by an adaptive resonance system,” Neural Networks, vol. 4, no. 6, pp. 759-771, May 1991.
[13] G. A. Carpenter, S. Grossberg, N. Markuzon, J. H. Reynolds, and D. B. Rosen,“Fuzzy ARTMAP: A neural network architecture for incremental supervised
learning of analog multidimensional maps,” IEEE Transactions on Neural Networks, vol. 3, no. 5, pp. 698-713, April 1992.
[14] C. Casetti, J. Kurose, and D. Towsley, “An adaptive algorithm for measurement-based admission control in integrated services packet networks,” Computer Communications, vol. 23, no. 14-15, pp. 1363-1376, August 2000.
[15] A. Charny, “Time scale analysis and scalability issues for explicit rate allocation in ATM networks,” IEEE/ACM Transactions on Networking, vol. 4, no. 4, pp. 569-581, August 1996.
[16] S. Chen, C. F. N. Cowan, and P. M. Grant, “Orthogonal least squares learning algorithm for radial basis function networks,” IEEE Transactions on Neural Networks, vol. 2, no. 2, pp. 302-309, March 1991.
[17] R. G. Cheng and C. J. Chang, “Design of a fuzzy traffic controller for ATM
networks,” IEEE/ACM Transactions on Networking, vol. 4, no. 3, pp. 460-469, November 1996.
[18] K. J. Cios and N. Liu, “A machine learning method for generation of a neural network architecture: A continuous ID3 algorithm,” IEEE Transactions on
Neural Networks, vol. 3, no. 2, pp. 280-290, May 1992.
[19] Y. L. Dae, M. K. Byung and S. C. Hyung, “A self-organized RBF network
combined with ART II,” Proceedings of IEEE International Joint Conference on
Neural Networks, vol. 3, pp. 1963-1968, July 1999.
[20] B. T. Doshi, “Deterministic rule based traffic descriptors for broadband ISDN: worst case behavior and connection acceptance control,” Proceedings Globecom, International Teletraffic Congress, ITC-14, vol. 1a, pp. 591-600, Antibes, France, June 1994.
[21] A. Elwalid, D. Mitra, and R. Wentworth, “A new approach for allocating buffers and bandwidth to heterogeneous regulated traffic in an ATM node,” IEEE J. Selected Areas in Communications, vol. 13, no. 6, pp. 1115-1127, August 1995.
[22] Z. Fan and P. Mars, “Fast and accurate estimation of ATM quality-of-service parameters with applications to call admission control,” IEE Proceedings Communications, vol. 145, no. 3, pp. 133-140, June 1998.
[23] S. E. Fahlman and C. Lebiere, “The cascade-correlation learning architecture,”Technical Report CMU-CS-90-100, School of Computer Science, Carnegie Mellon University, 1990.
[24] T. Frank, K. F. Kraiss, and T. Kuhlen, “Comparative analysis of fuzzy ART
and ART-2A network clustering performance,” IEEE Transactions on Neural
Networks, vol. 9, no. 3, pp. 544-559, May 1998.
[25] S. Geva and K. Malmstrom, “Local cluster neural net: Architecture, training, and applications,” Neurocomputing, vol. 20, no. 1, pp. 36-56, March 1998.
[26] R. M. Goodman, C. M. Higgins and J. W. Miller, “Rule-based neural network
for classification and probability estimation,” Neural Computation, vol. 4, no. 6, pp. 781-804, May 1992.
[27] R. Guerin, H. Ahmadi, and M. Nagshineh, “Equivalent capacity and its application to bandwidth allocation in high-speed networks,” IEEE J. Selected Areas in Communications, vol. 9, no. 7, pp. 968-981, September 1991.
[28] F.M. Ham and I. Kostanic, Principles of Neurocomputing for Science and Engineering, McGraw-Hill International, Singapore, 2000.
[29] Q. Hu and D. W. Petr, “Self-tuning fuzzy traffic rate control network,” Proceedings of IEEE International Conference on Communications, Dallas, Tx, pp. 424-428, June 1996.
[30] ITU-T, “Traffic Control and Congestion Control in B-ISDN,” Draft Recommendations, 1995.
[31] L. C. Jain and V. R. Vemuri, Industrial Applications of Neural Networks, CRC Press, New York, 1999.
[32] J. S. Jang, “ANFIS: Adaptive-network-based fuzzy inference systems,” IEEE
Transactions on Systems, Man, and Cybernetics, vol. 23, no. 3, pp. 665-685, May 1993.
[33] Y. Kawamura and H. Saito, “VP bandwidth management with dynamic connection admission control in ATM networks,”in Local Metropolitan Communication Systems, T. Hasegawa et al., Eds., London, U.K.: Chapman & Hall, vol. 3, 1995.
[34] K. Kobayashi, “Steady-state analysis of ATM multiplexer with variable input rate through diffusion approximation,” Performance Evaluation, vol. 23, no. 2, pp. 163-184, May 1995.
[35] H. Kobayashi and Q. Ren, “A diffusion approximation analysis of an ATM statistical multiplexer with multiple types of tra c, Part I: Equilibrium state solutions,”in Proceedings of IEEE International Conference on Communications, vol. 2, pp. 1047-1053, May 1993.
[36] S. Y. Kung, Digital Neural Networks, Prentice Hall, International, Inc., New York, 1993.
[37] S. Y. Kung and J. N. Hwang, “An algebraic projection analysis for optimal
hidden units size and learning rate in back-propagation learning,” Proceedings
of IEEE International Conference on Neural Networks, vol. 1, pp. 363-370, San
Diego, July 1988.
[38] W. G. Lai, “A flow control scheme on ATM networks with max-min fairness,”
Technical Report, National Science Council, Taiwan, 1997.
[39] Y. C. Lai and S. F. Tsai, “Some fair measurement-based admission controls,”in Proceedings of IEEE Globecom, vol. 4, pp. 2566-2570, San Antonio, Texas, November 2001.
[40] S. Lee and J. Song, “An integrated call admission control in ATM networks,”IEICE Transactions on Communications, vol. E82-B, no. 5, pp. 704-711, May 1999.
[41] T. K. Lee, M. Zukerman, and R. G. Addie, “Admission control schemes for bursty multimedia traffic,” in Proceedings of IEEE INFOCOM, vol. 1, pp. 478-487, Anchorage, Alaska, April 2001.
[42] S.-J. Lee, M.-T. Jone and H.-L. Tsai, “Constructing neural networks for multi-class discretization based on information entropy,” IEEE Transactions on Systems, Man,
and Cybernetics, vol. 29, no. 3, pp. 445-453, June 1999.
[43] S.-J. Lee and C.-L. Hou, “A neural-fuzzy system for congestion control in ATM networks,” IEEE Transactions on Systems, Man, and Cybernetics, vol. 30, no. 1, pp. 2-9, February 2000.
[44] S.-J. Lee and C.-L. Hou, “An ART-Based Construction of RBF Networks,” IEEE Transactions on Neural Networks, vol. 13, no. 6, pp. 1308-1321, November. 2002.
[45] C. C. Lee and P. C. Chung, “Robust radial basis function neural networks,”IEEE Transactions on Systems, Man, and Cybernetics - Part B: Cybernetics, vol. 29, no. 6, pp. 674-685, December 1999.
[46] A. Leonardis and H. Bischof, “An effcient MDL-based construction of RBF
networks,” Neural Networks, vol. 11, no. 5, pp. 963-973, July 1998.
[47] C. J. Lin and C. T. Lin, “Reinforcement learning for an ART-based fuzzy adaptive learning control network,” IEEE Transactions on Neural Networks, vol. 7, no. 3, pp. 709-731, May 1996.
[48] D. Liu and Y. Cai, “A heuristic approach for measurement-based admission
control with variable-size window,” in Proceedings of IEEE Globecom, vol. 4, pp. 2537-2541, San Antonio, Texas, November 2001.
[49] Y. C. Liu, “Rate regulation with feedback controller in ATM network - A neural network approach,” IEEE Journal on Selected Area in Communciations, vol. 15, no. 2, pp. 200-208, February 1997.
[50] Y. C. Liu and C. Douligeris, “Rate regulation with feedback controller in ATM networks: A neural network approach,” IEEE Journal on Selected Area in Communications, vol. 15, no. 2, February 1997.
[51] C. J. Lin and C. T. Lin, “An ART-based fuzzy adaptive learning control network,” IEEE Transactions on Fuzzy Systems, vol. 5, no. 4, pp. 477-496, November 1997.
[52] J. P. Nadal, “New algorithms for feedforward networks,” Neural Networks and SPIN Classes, Theumann and Koberle, Eds., pp 80-88, World Scienti c, Singapore, 1989.
[53] Nelson O.L. Ng and C. K. Tham, “Connection admission control of ATM network using integrated MLP and fuzzy controllers,” Computer Networks, vol. 32, no. 1, pp. 61-79, January 2000.
[54] N. Mark and G. Ramamurthy, “Real-time estimation of UPC parameters for
arbitrary traffic sources in ATM networks,” IEEE/ACM Transactions on Networking, vol. 6, no. 6, pp. 811-827, December 1998.
[55] J. L. McClelland and D. E. Rumehart, Parallel Distributed Processing (Two
Volumes), MIT Press, Cambridge, MA, 1986.
[56] OPNET Technologies, Inc. OPNET Modeler 7.0.B. Bethesda, MD, June 2000.
[57] M. J. L. Orr, “Regularization in the selection of radial basis function centers,”Neural Computation, vol. 7, no. 3, pp. 606-623, April 1995.
[58] J. Qiu and E. W. Knightly, “Measurement-Based Admission Control with Aggregate, Traffic Envelopes,” IEEE/ACM Transactions on Networking, vol. 9, no. 2, pp. 199-210, April 2001.
[59] A. Raha, W. Zhao, S. Kamat, and W. Jia, “Admission control for hard real-time connections in ATM LANs,” IEE Proceedings-Communications, vol. 148, no. 4, pp. 217-228, August 2001.
[60] G. Ramamurthy and Q. Ren, “Multiclass connection admission control policy
for ATM switches,” in Proceedings of IEEE INFOCOM'97, April 1997.
[61] Q. Ren and H. Kobayashi, “Diffusion approximation modeling for Markov modulated bursty traffic and its applications to bandwidth allocation in ATM networks,” IEEE J. Selected Areas in Communications, vol. 16, no. 2, pp. 679-691, June 1998.
[62] Q. Ren and G. Ramamurthy, “A real-time dynamic connection admission controller based on traffic modeling, measurement, and fuzzy logic control,” IEEE J. Selected Areas in Communications, vol. 18, no. 2, pp. 184-196, February 2000.
[63] L. Roberts, “Enchanced PRCA (proportional rate-control algorithm),” Technical Report AF-TM, pp. 154-157, 1994.
[64] H. Saito and K. Shiomoto, “Dynamic call admission control in ATM networks,”IEEE J. Selected Areas in Communications, vol. 9, no. 7, pp. 982-989, September 1991.
[65] P. K. Simpson, “Fuzzy min-max neural networks-Part 2: Clustering,” IEEE
Transactions on Fuzzy System, vol. 1, no. 1, pp. 32-45, February 1993.
[66] A. A. Tar and I. W. Habib, “Congestion control mechanism for ATM networks
using neural network,”in Proceedings of IEEE International Conference on Communications, pp. 206-210, June 1995.
[67] A. Tarraf, I. Habib, and T. Saadawi , “A neurocomputing approach to congestion control in an ATM multiplexer,” Journal on High Speed Networks, vol. 145, no. 4, pp. 329-346, January 1996.
[68] G. Tontini and A. A. de Queiroz, “RBF FUZZY-ARTMAP: a new fuzzy neural
network for robust on-line learning and identi cation of patterns,”IEEE International Conference on Systems, Man and Cybernetics, vol. 2, pp. 1364-1369, October 1996.
[69] G.-Q. Wei and G. Hirzinger, “Parametric shape-from-shading by radial basis
function,” IEEE Transactions on Pattern Analysis and Machine Intelligence,
vol. 19, no. 4, pp. 353-364, April 1997.
[70] S. Yan and M. Masaharu, “A new approach of neuro-fuzzy learning algorithm
for tuning fuzzy rules,” Fuzzy Sets and Systems, vol. 112, no. 16, pp. 99-116, May 2000.
[71] L. Yingwei, N. Sundararajan, and P. Saratchandran, “Performance evaluation
of a sequential minimal radial basis function (RBF) neural network learning
algorithm,” IEEE Transactions on Neural Networks, vol. 9, no. 2, pp. 308-318, March 1998.
[72] Y. Q. Zhang and A. Kandel, “Compensatory Neurofuzzy Systems with Fast
Learning Algorithms,” IEEE Transactions on Neural Networks, vol. 9, no. 1,
pp. 83-105, January 1998.
[73] F. Zhao, M. Hanawa, and M. Takahara, “Multiple-criteria call admission control scheme for ATM networks,” IEE Proceedings-Communications, vol. 148, no. 4, pp. 175-180, August 2001.
[74] Q. Zhu and L. Liu, “A global learning algorithm for a RBF network,” Neural Networks, vol. 12, no. 3, pp. 527-540, December 1999.