透過電腦模擬而得到一系統的反應曲面(Response Surface)，通常可以幫助我們了解系統的複雜動態關係。然而，此一反應曲面(Response Surface)(如存貨及延遲時間等)通常是許多控制參數(如控制策略(Control policy)、輸入值(Inputs)與服務時間的分配(Distribution)及佇列大小(Buffer Size)等)的複雜函數。而且較大的系統進行電腦模擬時所需的時間和運用的資源相對也較多。因此，如何選取用於電腦模擬的輸入點是一項相當重要的問題；選取好的輸入點可以充分地描繪系統的反應曲面，同時也能使電腦模擬的成本降到最低。
在本篇論文中，我們將設計一個模擬複雜網路系統的高效率演算法。我們所要提出的方法結合了分類迴歸樹(CART,即Classification And Regression Trees)及實驗設計中的D最佳化設計準則(D-Optimal Design Criteria)和U最佳化設計準則(U-Optimal Design Criteria)；我們也將此方法應用於一廣泛的網路模型。

None

[ 1] Hajek, B. “Hitting-time and Occupation-time Bounds Implied By Drift Analysis with Applications”, Annals of Applied Probability. 14, pp.501-525
[ 2] Pemantle, R. ; Rosenthal, J. S. “Moment conditions for a sequence with negative drift to be uniformly bounded in Lr ”, Stochastic Processes and their Applications 82, pp.143-155 (1999)
[ 3] Dai, J. G. “On positive Harris recurrence of multiclass queueing networks: a unified approach via fluid limit models”, Annals of Applied Probability, Vol. 5, pp.49-77, (1995)
[ 4] J. G. Dai and G. Weiss, "Stability and Instability of Fluid Models for Re-Entrant Lines",Mathematics of Operations Research, Vol. 21 , 115-134 (1996)
[ 5] Muhammad E1-Taha,University of Southern Maine, Sample-Path Analisis of Queueing Systems ,Portland,USA and Shaler Stidham Jr. (1998)
[ 6] Neter,Kutner,Nachtsheim,Wasserman 著,陳立信編譯,2000,實驗設計
最佳反應曲面設計 p358~p390
[ 7] Federov, V.V.(1972),Theory of Optimal Experiments,Academic Press,New York
[ 8] Breiman,L.,Friedman,J.H.,Olshen,R. A. &Stone,C.J.,Classification and Regression Trees, Monterey:Wadsworth(1984)
[ 9] W.N. Venables B.D.Ripley Modern Applied Statistics with S-plus Third Edition‘Tree-based methods’p303-p327
[10] Fang, K.T., Lin, D.K.J., Winker, P. and Zhang, Y. , Uniform design: Theory and applications, Technometrics, in press(2000)
[11] Hung,Y.C.,Modeling and Analysis of stochastic Networks with Shared
Resources,Ph.D.thesis,Department of Statistics,The University of Michigan(2002)
[12] Hung, Y. C., Michailidis, G., Bingham, D. R.. "Developing Efficient Simulation Methodology for Complex Queueing Networks", Proceedings of the Winter Simulation Conference 2003, New Orleans, pp. 152-159
[13] Friedman,J.,Multivariate adaptive regression splines,Annals ofStatistics,19,1-14(1991)
[14http://gsbwww.uchicago.edu/computing/research/SASManual/qc/chap24/sect30.htm,Johnson, Moore, and Ylvisaker (1990)
[15] Sacks, J., Welch, W.J., Mitchell, T.J.&Wynn, H.P., Design and analysis of computer experiments, Statistical Science, 4,409-423(1989)
[16] Box,G.E.P.&Draper,D.R., Empirical Model Building and Response Surface, New York:John Wiley &Sins(1987)
[17] Xie, M.Y. and Fang, K.T., Admissibility and minimaxity of the uniform design measure in nonparametric regression model, J. Statist., Planning and Inference, 83, 101-111(1999)
[18] Harries, Steve, Networking and Telecommunications for Information Systems: AnIntroduction to Information Networking. London: Library Association
Publishing,(1993)
[19] C.Cortes and V.Vapnik. Support vector networks. Machine Learning,20: 273—297,(1995)
[20] Vladimir N. Vapnik, "The Nature of Statistical Learning Theory", ISBN0-387-94559-8,Springer-Verlag, (1995)
[21] Wu, C.F.J. and Hamada, M. "Experiments:Planning, Analysis, and Parameter Design", New York: Wiley,(2000)
[22] W.N. Venables B.D.Ripley Modern Applied Statistics with S-plus Third Edition‘Tree-based methods’p281-p302