Quadratic Convergence and Numerical Experiments of Newton ’ s Method for Computing the Nearest Correlation Matrix

The nearest correlation matrix problem is to find a correlation matrix which is closest to a given symmetric matrix under the Frobenius norm. The well studied dual approach is to reformulate this problem as an unconstrained continuously differentiable convex optimization problem. Gradient methods and quasi-Newton methods like BFGS have been used directly to obtain globally convergent methods. Since the objective function in the dual approach is not twice continuously differentiable, these methods converge at best linearly. In this paper, we investigate a Newton-type method for the nearest correlation matrix problem. Based on recent developments on strongly semismooth matrix valued functions, we prove the quadratic convergence of the proposed Newton’s method. Numerical experiments confirm the fast convergence and the high efficiency of the method. AMS subject classifications. 49M45, 90C25, 90C33

[1]  Frank Deutsch,et al.  A Dual Approach to Constrained Interpolationfrom a Convex Subset of Hilbert Space , 1997 .

[2]  Jonathan M. Borwein,et al.  Partially finite convex programming, Part I: Quasi relative interiors and duality theory , 1992, Math. Program..

[3]  W. Pollul,et al.  Convex interpolation with minimalL∞-norm of the second derivative , 1984 .

[4]  Paul Tseng,et al.  Merit functions for semi-definite complemetarity problems , 1998, Math. Program..

[5]  L. Andersson,et al.  Am algorithm for constrained interpolation , 1987 .

[6]  N. Higham.  Computing the nearest correlation matrix—a problem from finance , 2002 .

[7]  Defeng Sun,et al.  Semismooth Matrix-Valued Functions , 2002, Math. Oper. Res..

[8]  C. Micchelli,et al.  Smoothing and Interpolation in a Convex Subset of a Hilbert Space , 1988 .

[9]  Paul Tseng,et al.  Bounded Linear Regularity, Strong CHIP, and CHIP are Distinct Properties , 2000 .

[10]  Heinz H. Bauschke,et al.  Strong conical hull intersection property, bounded linear regularity, Jameson’s property (G), and error bounds in convex optimization , 1999, Math. Program..

[11]  Hans-Peter Seidel,et al.  Local compliance estimation via positive semidefinite constrained least squares , 2004, IEEE Transactions on Robotics.

[12]  R. Mifflin.  Semismooth and Semiconvex Functions in Constrained Optimization , 1977 .

[13]  M. Hestenes,et al.  Methods of conjugate gradients for solving linear systems , 1952 .

[14]  Liqun Qi,et al.  A nonsmooth version of Newton's method , 1993, Math. Program..

[15]  Karl Löwner.  Über monotone Matrixfunktionen , 1934 .

[16]  Asen L. Dontchev,et al.  Duality and well-posedness in convex interpolation ∗) , 1989 .

[17]  Defeng Sun,et al.  A further result on an implicit function theorem for locally Lipschitz functions , 2001, Oper. Res. Lett..

[18]  John E. Dennis,et al.  Numerical methods for unconstrained optimization and nonlinear equations , 1983, Prentice Hall series in computational mathematics.

[19]  Charles K. Chui,et al.  Constrained best approximation in Hilbert space , 1990 .

[20]  L. D. Irvine,et al.  Constrained interpolation and smoothing , 1986 .

[21]  W. Donoghue.  Monotone Matrix Functions and Analytic Continuation , 1974 .

[22]  D. Luenberger.  Optimization by Vector Space Methods , 1968 .

[23]  Jérôme Malick,et al.  A Dual Approach to Semidefinite Least-Squares Problems , 2004, SIAM J. Matrix Anal. Appl..

[24]  R. Dykstra.  An Algorithm for Restricted Least Squares Regression , 1983 .

[25]  Stephen P. Boyd,et al.  Least-Squares Covariance Matrix Adjustment , 2005, SIAM J. Matrix Anal. Appl..

[26]  M. Kojima,et al.  EXTENSION OF NEWTON AND QUASI-NEWTON METHODS TO SYSTEMS OF PC^1 EQUATIONS , 1986 .

[27]  U. Hornung,et al.  Interpolation by smooth functions under restrictions on the derivatives , 1980 .

[28]  Paul Tseng,et al.  Analysis of Nonsmooth Symmetric-Matrix-Valued Functions with Applications to Semidefinite Complementarity Problems , 2003, SIAM J. Optim..

[29]  N. Higham.  COMPUTING A NEAREST SYMMETRIC POSITIVE SEMIDEFINITE MATRIX , 1988 .

[30]  Wu Li,et al.  Best Approximation from the Intersection of a Closed Convex Set and a Polyhedron in Hilbert Space, Weak Slater Conditions, and the Strong Conical Hull Intersection Property , 1999, SIAM J. Optim..

[31]  F. Clarke.  Optimization And Nonsmooth Analysis , 1983 .

[32]  Francisco Facchinei,et al.  Minimization of SC1 functions and the Maratos effect , 1995, Oper. Res. Lett..

[33]  Liqun Qi,et al.  Convergence of Newton's method for convex best interpolation , 2001, Numerische Mathematik.

[34]  Qi,et al.  Quadratic Convergence of Newton's Method for Convex Interpolation and Smoothing , 2003 .

[35]  Charles K. Chui,et al.  Constrained best approximation in Hilbert space, II , 1992 .

[36]  J. Pang,et al.  A globally convergent Newton method for convex SC1 minimization problems , 1995 .