在本論文中，我們提出求最大結構奇異值上界的演算法，此演算法在μ-合成控制問題裡扮演著重要角色。如文獻所記載，求最大結構的奇異值是 NP-hard之問題，因此我們在這篇論文中專注於求最大結構奇異值之上界。我們發展出一個牛頓型態之演算法，也推導一些和演算法有關的理論。從數值結果說明這新開發出來的演算法是個有效的演算法。

In this dissertation, numerical algorithms for the computation of an upper bound of the largest structured singular value which always plays a key role to the μ-synthesis control problem are proposed. As shown in literatures the computations for the largest structured singular value is an NP-hard problem. We hence concentrate on the study for computing an upper bound of the largest structured singular value in this dissertation. A Newton’s type algorithm is developed. Some related theoretical results of the algorithm are investigated. Numerical implementation shows the efficiency of the new developed algorithm.

[1] A. L. Andrew, K. W. E. Chu, and P. Lancaster, Derivatives of eigenvalues and eigenvectors of matrix function, SIAM J. Matrix Anal. Appl.,(1993), Vol. 14, No. 4, pp. 903−926.
[2] G. J. Balas, J. C. Doyle, K. Glover, A. Packard, and R. Smith, μ-Analysis and Synthesis Toolbox, Math Works, Natick, MA, 1995, Chap. 4.
[3] S. P. Boyd and V. Balakrishnan, A regularity result for the singular values of a transfer matrix and a quadratically convergent algorithm for computing its L1-norm, Systems and Control Letters, (1990), Vol. 15, pp. 1−7.
[4] S. P. Boyd, L. Ghaoui, E. Feron, and V. Balakrishnan, Linear Matrix Inequalities in System and Control Theory, SIAM, Philadelphia, 1994.
[5] R. Braatz, P. Young, J. Doyle, and M. Morari, Computational complexity of μ calculation, IEEE Transactions on Automatic Control, (1994), Vol. AC-39, No. 5, pp. 1000−1002.
[6] J. Br¨auninger, A quasi-Newton method with Cholesky factorization, Computing, (1980), Vol. 25, pp. 155−162.
[7] R. H. Chan, S. F. Xu, and H. M. Zhou, On the convergence rate of a quasi-newton method for inverse eigenvalue problems, SIAM J. Numer. Anal. (1999), Vol. 36, No. 2, pp. 436−441.
[8] W. Cheney and D. Kincaid, Numerical Mathematics and Computing, Brooks/Cole, 1994.
[9] J. C. Doyle, Analysis of feedback system with structured uncertainties, IEEE Proc., Part D, (1982), Vol. 129, No. 6, pp. 242−250.
[10] M. Fan, A. Tits, and J. Doyle, Robustness in the presence of mixed parametric uncertainty and unmodelled dynamics, IEEE Transactions on Automatic Control, (1991), Vol. AC-36, No. 1, pp. 25−38.
[11] S. C. Fang and S. Puthenpura, Linear Optimization and Extensions. Prentice Hall, N. J., 1993.
[12] J. A. Ford and I. A. Moghrabi, Alternative parameter choices for multistep quasi-newton methods, Optim. Meth. Software, (1993), Vol. 2, pp. 357−370.
[13] J. A. Ford and I. A. Moghrabi, Multi-step quasi-newton methods for optimization, J. Comput. Appl. Math., (1994), Vol. 50, pp. 305−323.
[14] J. A. Ford and I. A. Moghrabi, Minimum curvature multi-step quasinewton methods, Computers Math. Applic., (1996), Vol. 31, pp. 179−186.
[15] G. H. Golub and C. F. Van Loan, Matrix Computations, 3rd ed., Johns Hopkins U. P., 1996.
[16] M. J. Hays, D. G. Bates, and I. Postlethwaite, New tools for computing tight bounds on the real structured singular value, Journal of Guidance, Control, and Dynamics, (2001), Vol. 24, No. 6, pp.1204−1213.
[17] H. N. Huang, M. L. Jiang, F. B. Yeh, and L. C. Fu, μ-synthesis via spectral interpolation theory, Proc. of 2003 Chinese Automatic Control Conference and Bio-Mechatronics System Control and Application Workshop, Chung-Li, Taiwan, 2003, pp. 1809−1814.
[18] C. T. Lawrence, A. L. Tits, and P. Van Dooren, A fast algorithm for the computation of an upper bound on the μ-norm, Automatica, (2000), Vol. 36, pp. 449−456.
[19] J. Lee and T. F. Edgar, Upper bounds of structured singular values for mixed uncertainties, Proc. of the 42nd IEEE Conference on Decision and Control, Maui, Hawaii USA, 2003, pp. 798−802.
[20] J. F. Magni and C. D¨oll, μ-analysis for flexible systems, ONERA-CERT, D´epartement d’Automatique B. P. 4025, F31055 Toulouse Cedex 04.
[21] A. Megretski, On the gap between structured singular values and their upper bounds , Proc. of the 32nd Conference on Decision and Control, San Antonio, Texas, 1993, pp. 3461−3462.
[22] M. L. Overton and P. Van Dooren, On computing the complex passivity radius. Proc. of the 44th IEEE Conference on Decision and Control and European Control Conference, Seville, Spain, 2005, pp. 7960−7964.
[23] A. Packard and J. Doyle, The complex structured singular value, Automatica, (1993), Vol. 29, no. 1, pp. 72−109.
[24] A. Packard and J. Doyle, A Power method for the structured singular value, Proc. of the American Control Conference, Vol. 2, Inst. of Electrical and Electronics Engineers, New York, 1988, pp. 1213−1218.
[25] M. Safonov and P. Lee, A multiplier method for computing real multivariable stability margins, Proc. of the IFAC Word Congress, Vol. 1, International Federation of Automatic Control, Oxford, 1993, pp. 275−278.
[26] T. F. Sturm, Quasi-Newton method by Hermit interpolation, J. of Optimization Theory and Applications, (1994), Vol. 83, No. 3, pp. 587−612.
[27] O. Toker and H. ¨Ozbay, On the complexity of purely complex μ computation and relate problems in multidimensional systems, IEEE Trans. Auto. Control, (1998), Vol. 43, no. 3, pp. 409−414.
[28] H. Z. Wang, Numerical Algorithms for Estimating Upper Bounds of μ-Norm, master thesis, NCKU, Tainan, Taiwan, 2004.
[29] P. M. Young, M. P. Newlin, and J. C. Doyle, Computing bounds for the mixed μ problem, International Journal of Robust and Nolinear Control, (1995), Vol. 5, No. 6, pp. 573−590.
[30] K. Zhou, J. C. Doyle, and K. Glover, Robust and Optimal Control, Prentice Hall, N.J., 1995.