Cholesky Decomposition

Work in progress. This note is still being written and incomplete.

2 min read Last updated Mon Aug 10 2026 06:34:55 GMT+0000 (Coordinated Universal Time)

A symmetric matrix AA is positive definite iff AA can be factored as

A=LLTA = LL^T

where LL is lower triangular with non-zero diagonal entries. This factorization is Cholesky decomposition. The 2 triangular factors are transposes of each other, so LL alone determines the factorization.

Computing LL

Multiplying out A=LLTA = LL^T and equating entries row by row gives, for row k=1,2,,nk = 1, 2, \ldots, n in order

lki=akij=1i1lijlkjlii,i=1,2,,k1l_{ki} = \frac{a_{ki} - \sum\limits_{j=1}^{i-1} l_{ij}l_{kj}}{l_{ii}}, \quad i = 1, 2, \ldots, k-1 lkk=akkj=1k1lkj2l_{kk} = \sqrt{a_{kk} - \sum\limits_{j=1}^{k-1} l_{kj}^2}

Example

Apply Cholesky decomposition to

A=(61555155522555225979)A = \begin{pmatrix} 6 & 15 & 55 \\ 15 & 55 & 225 \\ 55 & 225 & 979 \end{pmatrix}

Row 11

l11=62.449l_{11} = \sqrt{6} \approx 2.449

Row 22

l21=a21l11=152.4496.124,l22=a22l212=556.12424.183l_{21} = \frac{a_{21}}{l_{11}} = \frac{15}{2.449} \approx 6.124, \quad l_{22} = \sqrt{a_{22} - l_{21}^2} = \sqrt{55 - 6.124^2} \approx 4.183

Row 33

l31=a31l11=552.44922.454,l32=a32l31l21l22=22522.454×6.1244.18320.917l_{31} = \frac{a_{31}}{l_{11}} = \frac{55}{2.449} \approx 22.454, \quad l_{32} = \frac{a_{32} - l_{31}l_{21}}{l_{22}} = \frac{225 - 22.454 \times 6.124}{4.183} \approx 20.917 l33=a33l312l322=97922.454220.91726.110l_{33} = \sqrt{a_{33} - l_{31}^2 - l_{32}^2} = \sqrt{979 - 22.454^2 - 20.917^2} \approx 6.110 L(2.449006.1244.183022.45420.9176.110)L \approx \begin{pmatrix} 2.449 & 0 & 0 \\ 6.124 & 4.183 & 0 \\ 22.454 & 20.917 & 6.110 \end{pmatrix}
Was this helpful?