The scaling and squaring method for the matrix exponential revisited

被引:369
作者
Higham, NJ [1 ]
机构
[1] Univ Manchester, Sch Math, Manchester M60 1QD, Lancs, England
关键词
matrix function; matrix exponential; Pade approximation; matrix polynomial evaluation; scaling and squaring method; MATLAB; expm; backward error analysis; performance profile;
D O I
10.1137/04061101X
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
The scaling and squaring method is the most widely used method for computing the matrix exponential, not least because it is the method implemented in MATLAB's expm function. The method scales the matrix by a power of 2 to reduce the norm to order 1, computes a Pade approximant to the matrix exponential, and then repeatedly squares to undo the effect of the scaling. We give a new backward error analysis of the method ( in exact arithmetic) that employs sharp bounds for the truncation errors and leads to an implementation of essentially optimal efficiency. We also give new rounding error analysis that shows the computed Pade approximant of the scaled matrix to be highly accurate. For IEEE double precision arithmetic the best choice of degree of Pade approximant turns out to be 13, rather than the 6 or 8 used by previous authors. Our implementation of the scaling and squaring method always requires at least two fewer matrix multiplications than expm when the matrix norm exceeds 1, which can amount to a 37% saving in the number of multiplications, and it is typically more accurate, owing to the fewer required squarings. We also investigate a different scaling and squaring algorithm proposed by Najfeld and Havel that employs a Pade approximation to the function x coth( x). This method is found to be essentially a variation of the standard one with weaker supporting error analysis.
引用
收藏
页码:1179 / 1193
页数:15
相关论文
共 24 条
[1]   A Schur-Parlett algorithm for computing matrix functions [J].
Davies, PI ;
Higham, NJ .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 2003, 25 (02) :464-485
[2]   Orthogonal eigenvectors and relative gaps [J].
Dhillon, IS ;
Parlett, BN .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 2003, 25 (03) :858-899
[3]   Pade approximation for the exponential of a block triangular matrix [J].
Dieci, L ;
Papini, A .
LINEAR ALGEBRA AND ITS APPLICATIONS, 2000, 308 (1-3) :183-202
[4]   Benchmarking optimization software with performance profiles [J].
Dolan, ED ;
Moré, JJ .
MATHEMATICAL PROGRAMMING, 2002, 91 (02) :201-213
[5]  
Franklin G.F., 1998, Digital Control of Dynamics Systems, V3rd ed.
[6]  
Gautschi W., 1997, NUMERICAL ANAL INTRO
[7]  
Golub G. H., 1996, MATRIX COMPUTATIONS
[8]   MATRIX DECOMPOSITIONS OF 2-DIMENSIONAL NUCLEAR-MAGNETIC-RESONANCE SPECTRA [J].
HAVEL, TF ;
NAJFELD, I ;
YANG, JX .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 1994, 91 (17) :7962-7966
[9]  
Higham N. J., 2002, ACCURACY STABILITY N
[10]  
Higham N.J., The Matrix Computation Toolbox