Chapter 1 Flashcards
Norm
A real number ||x|| such that:
i) ||x||>0 , for all x=/=0
ii) ||αx||=|α| ||x||
iii) ||x+y||<=||x|| + ||y||
Subordinate Matrix Norm
||A|| = max (||Ax||/||x||)
i) ||A|| > 0, for all A=/=0
ii) ||αA|| = |α| ||A||
iii) ||Ax|| <= ||A|| ||x||
iv) ||AB|| <= ||A|| ||B||
v) ||A + B|| <= ||A|| + ||B||
Cauchy-Schwartz Inequality
² <=
P-norm
||x|| = (∑|xi|^p)^(1/p)
Infinity - norm
||x|| = max (|xi|)
Spectral radius
p(A) = max (λ)
p(A) <= ||A||
Estimating B inverse
<= 1/(1-||A||)
where B = I + A and ||A||<1
Error in x when the system is changed by δb
<= K(A)||δb||/||b||
Condition Number
K(A)=||A|| ||A-1||
K(A) >= 1
K(A) >= |λ1|/|λn|
K(αA) = K(A)
Matrix norm subordinate to the infinity norm
Maximum row sum
Matrix norm subordinate to the L1 normed vector space
Maximum column sum
||x+y||² <= ||x||² + ||y||²
||x+y||² = = + 2 + <= + + 2sqrt() by C-S <= + = ||x||² + ||y||²