GRADIENT ALGORITHMS for COMMON LYAPUNOV FUNCTIONS Daniel Liberzon Univ. of Illinois at Urbana-Champaign, U.S.A. Roberto Tempo IEIIT-CNR, Politecnico di.

Slides:



Advertisements
Similar presentations
COMMON WEAK LYAPUNOV FUNCTIONS and OBSERVABILITY Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of Illinois.
Advertisements

1 STABILITY OF SWITCHED SYSTEMS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of Illinois at Urbana-Champaign.
CONTROL with LIMITED INFORMATION Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of Illinois at Urbana-Champaign.
Statistical perturbation theory for spectral clustering Harrachov, 2007 A. Spence and Z. Stoyanov.
INPUT-TO-STATE STABILITY of SWITCHED SYSTEMS Debasish Chatterjee, Linh Vu, Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer.
TOWARDS ROBUST LIE-ALGEBRAIC STABILITY CONDITIONS for SWITCHED LINEAR SYSTEMS 49 th CDC, Atlanta, GA, Dec 2010 Daniel Liberzon Univ. of Illinois, Urbana-Champaign,
1 of 13 STABILIZING a SWITCHED LINEAR SYSTEM by SAMPLED - DATA QUANTIZED FEEDBACK 50 th CDC-ECC, Orlando, FL, Dec 2011, last talk in the program! Daniel.
1 of 14 LIMITED - INFORMATION CONTROL of SWITCHED and HYBRID SYSTEMS via PROPAGATION of REACHABLE SETS HSCC, Philadelphia, April 2013 Daniel Liberzon Coordinated.
CONTROL of SWITCHED SYSTEMS with LIMITED INFORMATION
Scientific Computing QR Factorization Part 2 – Algorithm to Find Eigenvalues.
Principal Component Analysis Based on L1-Norm Maximization Nojun Kwak IEEE Transactions on Pattern Analysis and Machine Intelligence, 2008.
Bregman Iterative Algorithms for L1 Minimization with
Neural and Evolutionary Computing - Lecture 4 1 Random Search Algorithms. Simulated Annealing Motivation Simple Random Search Algorithms Simulated Annealing.
INTRODUCTION to SWITCHED SYSTEMS ; STABILITY under ARBITRARY SWITCHING
THE ROLE OF LIE BRACKETS IN STABILITY OF LINEAR AND NONLINEAR SWITCHED SYSTEMS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical &
IFAC AIRTC, Budapest, October 2000 On the Dynamic Instability of a Class of Switching System Robert Noel Shorten Department of Computer Science National.
NONLINEAR OBSERVABILITY NOTIONS and STABILITY of SWITCHED SYSTEMS CDC ’02 João Hespanha Univ. of California at Santa Barbara Daniel Liberzon Univ. of Illinois.
U NIVERSITY OF M ASSACHUSETTS, A MHERST Department of Computer Science Solving POMDPs Using Quadratically Constrained Linear Programs Christopher Amato.
ACCELERATING GOOGLE’S PAGERANK Liz & Steve. Background  When a search query is entered in Google, the relevant results are returned to the user in an.
TOWARDS ROBUST LIE-ALGEBRAIC STABILITY CONDITIONS for SWITCHED LINEAR SYSTEMS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical &
Venkataramanan Balakrishnan Purdue University Applications of Convex Optimization in Systems and Control.
1 Nonlinear Control Design for LDIs via Convex Hull Quadratic Lyapunov Functions Tingshu Hu University of Massachusetts, Lowell.
1 of 9 ON ALMOST LYAPUNOV FUNCTIONS Daniel Liberzon University of Illinois, Urbana-Champaign, U.S.A. TexPoint fonts used in EMF. Read the TexPoint manual.
Lecture #13 Stability under slow switching & state-dependent switching João P. Hespanha University of California at Santa Barbara Hybrid Control and Switched.
STABILITY under CONSTRAINED SWITCHING Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of Illinois at Urbana-Champaign.
COMMUTATION RELATIONS and STABILITY of SWITCHED SYSTEMS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of.
IEIIT-CNR SWAN, ARRI ©RT Las Vegas and Monte Carlo Randomized Algorithms for Systems and Control Roberto Tempo IEIIT-CNR Politecnico di Torino
1cs542g-term Notes  In assignment 1, problem 2: smoothness = number of times differentiable.
Discovering Affine Equalities Using Random Interpretation Sumit Gulwani George Necula EECS Department University of California, Berkeley.
1cs542g-term Notes  Assignment 1 due tonight ( me by tomorrow morning)
Jie Gao Joint work with Amitabh Basu*, Joseph Mitchell, Girishkumar Stony Brook Distributed Localization using Noisy Distance and Angle Information.
Nonlinear Sampling. 2 Saturation in CCD sensors Dynamic range correction Optical devices High power amplifiers s(-t) Memoryless nonlinear distortion t=n.
Unconstrained Optimization Problem
Greg GrudicIntro AI1 Introduction to Artificial Intelligence CSCI 3202: The Perceptron Algorithm Greg Grudic.
Greg GrudicIntro AI1 Support Vector Machine (SVM) Classification Greg Grudic.
A LIE-ALGEBRAIC CONDITION for STABILITY of SWITCHED NONLINEAR SYSTEMS CDC ’04 Michael Margaliot Tel Aviv University, Israel Daniel Liberzon Univ. of Illinois.
1 of 12 COMMUTATORS, ROBUSTNESS, and STABILITY of SWITCHED LINEAR SYSTEMS SIAM Conference on Control & its Applications, Paris, July 2015 Daniel Liberzon.
CONTROL of NONLINEAR SYSTEMS with LIMITED INFORMATION Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of.
1 of 17 NORM - CONTROLLABILITY, or How a Nonlinear System Responds to Large Inputs Daniel Liberzon Univ. of Illinois at Urbana-Champaign, U.S.A. Workshop.
9 1 Performance Optimization. 9 2 Basic Optimization Algorithm p k - Search Direction  k - Learning Rate or.
MEETING THE NEED FOR ROBUSTIFIED NONLINEAR SYSTEM THEORY CONCEPTS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng.,
Algorithms for a large sparse nonlinear eigenvalue problem Yusaku Yamamoto Dept. of Computational Science & Engineering Nagoya University.
TUTORIAL on LOGIC-BASED CONTROL Part I: SWITCHED CONTROL SYSTEMS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng.,
QUANTIZED OUTPUT FEEDBACK CONTROL of NONLINEAR SYSTEMS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of.
Lecture #11 Stability of switched system: Arbitrary switching João P. Hespanha University of California at Santa Barbara Hybrid Control and Switched Systems.
Michael J. Neely, University of Southern California CISS, Princeton University, March 2012 Asynchronous Scheduling for.
Efficient computation of Robust Low-Rank Matrix Approximations in the Presence of Missing Data using the L 1 Norm Anders Eriksson and Anton van den Hengel.
Low-Rank Kernel Learning with Bregman Matrix Divergences Brian Kulis, Matyas A. Sustik and Inderjit S. Dhillon Journal of Machine Learning Research 10.
Daniel Liberzon Coordinated Science Laboratory and
Lyapunov exponents of products of free operators Vladislav Kargin (CIMS)
Biointelligence Laboratory, Seoul National University
COMMUTATION RELATIONS and STABILITY of SWITCHED SYSTEMS Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of.
Traditional Approaches to Modeling and Analysis
SMALL-GAIN APPROACH to STABILITY ANALYSIS of HYBRID SYSTEMS CDC ’05 Dragan Nešić University of Melbourne, Australia Daniel Liberzon Univ. of Illinois at.
NONLINEAR CONTROL with LIMITED INFORMATION Daniel Liberzon Coordinated Science Laboratory and Dept. of Electrical & Computer Eng., Univ. of Illinois at.
STABILITY of SWITCHED SYSTEMS – family of asymptotically stable systems – piecewise constant switching signal Want GUAS w.r.t. want GUES For switched linear.
Lecture #7 Stability and convergence of ODEs João P. Hespanha University of California at Santa Barbara Hybrid Control and Switched Systems NO CLASSES.
Giansalvo EXIN Cirrincione unit #4 Single-layer networks They directly compute linear discriminant functions using the TS without need of determining.
Generalization Error of pac Model  Let be a set of training examples chosen i.i.d. according to  Treat the generalization error as a r.v. depending on.
PROBABILITY AND COMPUTING RANDOMIZED ALGORITHMS AND PROBABILISTIC ANALYSIS CHAPTER 1 IWAMA and ITO Lab. M1 Sakaidani Hikaru 1.
Q-Learning for Policy Improvement in Network Routing
Computational Optimization
Generalization and adaptivity in stochastic convex optimization
Input-to-State Stability for Switched Systems
Probabilistic Models for Linear Regression
Jianping Fan Dept of CS UNC-Charlotte
A Dynamic System Analysis of Simultaneous Recurrent Neural Network
Guosong Yang1, A. James Schmidt2, and Daniel Liberzon2
Performance Optimization
On Topological Entropy and Stability of Switched Linear Systems
Presentation transcript:

GRADIENT ALGORITHMS for COMMON LYAPUNOV FUNCTIONS Daniel Liberzon Univ. of Illinois at Urbana-Champaign, U.S.A. Roberto Tempo IEIIT-CNR, Politecnico di Torino, Italy

PROBLEM Motivation: stability of uncertain and switched systems Analytical results: hard to come by (beyond ) require special structure LMI methods: can handle large finite families provide limited insight Our approach: gradient descent iterations handle inequalities sequentially Goal: algorithmic approach with theoretical insight Given Hurwitz matrices and matrix, find matrix :

MOTIVATING EXAMPLE quadratic common Lyapunov function In the special case when matrices commute: Nonlinear extensions: Shim et al. (1998), Vu & L (2003) (Narendra & Balakrishnan, 1994)

I TERATIVE A LGORITHMS : PRIOR WORK Algebraic inequalities: Agmon, Motzkin, Schoenberg (1954) Polyak (1964) Yakubovich (1966) Matrix inequalities: Polyak & Tempo (2001) Calafiore & Polyak (2001)

G RADIENT A LGORITHMS : PRELIMINARIES – convex differentiable real-valued functional on the space of symmetric matrices, Examples: (need this to be a simple eigenvalue) 1. ( is Frobenius norm, is projection onto matrices ) 2.

G RADIENT A LGORITHMS : PRELIMINARIES – convex differentiable real-valued functional on the space of symmetric matrices, Gradient: ( is unit eigenvector of with eigenvalue ) – convex in given

G RADIENT A LGORITHMS : DETERMINISTIC CASE – finite family of Hurwitz matrices – arbitrary symmetric matrix Gradient iteration: Theorem: Solution, if it exists, is found in a finite number of steps – visits each index times Idea of proof: distance from to solution set decreases at each correction step ( – suitably chosen stepsize)

G RADIENT A LGORITHMS : PROBABILISTIC CASE Idea of proof: still get closer with each correction step correction step is executed with prob. 1 – compact (possibly infinite) family – picked using probability distribution on s.t. every relatively open subset has positive measure Theorem:Solution, if it exists, is found in a finite number of steps with probability 1 Gradient iteration (randomized version):

SIMULATION EXAMPLE Interval family of triangular Hurwitz matrices: vertices Deterministic gradient: ( ineqs): 10,000 iterations (a few seconds) ( ineqs): 10,000,000 iterations (a few hours) Compare: quadstab ( MATLAB ) stacks when Randomized gradient gives faster convergence Both are quite easy to program

CONCLUSIONS Gradient iteration algorithms for solving simultaneous Lyapunov matrix inequalities Deterministic convergence for finite families, probabilistic convergence for infinite families Open issues: performance comparison of different methods addressing existence of solutions optimal schedule of iterations additional requirements on solution Contribution: