School address: University of California, San Diego Department of MAE 9500 Gilman Dr, m/c 0411 La Jolla, CA  92093-0411 (858) 822-3391 Home address: 6699 Beadnell Way, Apt. 210 San Diego, CA  92117 (858) 495-0289 Email: scott@turbulence.ucsd.edu

photo courtesy of Dennis Williams Photography

Research Interests

• Large-scale optimization, especially for linear matrix inequalities and semidefinite programs
• Nonsmooth optimization, bundle methods
• Control and estimation strategies for distributed systems
• Numerical linear algebra, especially eigenvalue/eigenvector computation

Thesis stuff

• Ph.D., Electrical and Computer Engineering, U.C. Santa Barabara, December 2001.
• Advisor: Roy Smith
• Title: An Inexact Bundle Method for Solving Large Structured Linear Matrix Inequalities.  [ps 2.8MB] [ps.gz 740kB] [pdf 1.4MB]

Publications & Presentations

• S. A. Miller.  Sensitivity of Solutions to Semidefinite Programs.  Center for Control Engineering and Computation Technical Report CCEC-97-0519, ECE Dept., Univ. of Calif., Santa Barbara, CA, May 19, 1997.  [postscript] [pdf]
• S. A. Miller and R. S. Smith.  A bundle method for efficiently solving large structured linear matrix inequalities.  In Proc. Amer. Contr. Conf., pages 1405-1409, Chicago, IL, June 2000.  [postscript] [pdf]
• S. A. Miller.  Efficient Computation of the Maximum Eigenvalue of Large Symmetric Matrices. Center for Control Engineering and Computation Technical Report CCEC-00-0817, ECE Dept., Univ. of Calif., Santa Barbara, CA, August 17, 2000.  [postscript] [pdf]
• S. A. Miller and R. S. Smith.  Solving large structured semidefinite programs using an inexact spectral bundle method.  In Proc. IEEE Conf. Dec. Contr., pages 5027-5032, Sydney, Australia, December 2000.  [postscript] [pdf]
• R. S. Smith, G. Dullerud and S. A. Miller.  Model validation for nonlinear systems.  In Proc. IEEE Conf. Dec. Contr., pages 1232-1236, Sydney, Australia, December 2000.  [pdf]
• S. A. Miller and R. S. Smith.  Efficient Convergence Test for the Maximum Eigenvalue Estimate from the Lanczos Method.  Submitted to SIAM J. Matrix Anal. Appl., August 2001.  [postscript] [pdf]
• S. A. Miller.  Solving Large Scale Linear Matrix Inequalities with a Reduced Bundle Method.  Presented at Optimization and Applications, Mathematisches Forschungsinstitut Oberwolfach, Germany, January 2002.  [postscript] [pdf]
• S. A. Miller.  An Interior Point Approach to Electronic Structure Computation.  Presented at Séminaire de Calcul Scientifique, CERMICS, ENPC, Champs sur Marne, France, June 2002.  [postscript] [pdf]
• S. A. Miller, T. R. Bewley and P. Cathalifaud.  Decentralized Optimal Estimation of Boundary Layer Flow Systems.  Presented at the Sixth Southern California Nonlinear Controls Workshop, U.C. San Diego, May 2003.  [gzipped postscript] [pdf]
• S. A. Miller and T. R. Bewley.  Multidimensional Control Design in Continuous Space, with Applications to Control of Channel Flow.  Submitted to CDC 2004, March 2004.  [postscript] [pdf]
• S. A. Miller and J. Malick.  Newton Methods for Nonsmooth Convex Minimization: Connections between U-Lagrangian, Riemannian Newton and SQP Methods.  Submitted to Math. Prog., May 2004.  [postscript] [pdf]   See also INRIA Rapport de Recherche no. RR-5185.