Class Notes (835,324)
Statistics (248)
STAT312 (22)
Lecture

Block matrices II, LU decomposition.pdf

5 Pages
102 Views

School
Department
Statistics
Course
STAT312
Professor
Douglas Wiens
Semester
Fall

Description
61 11. Block matrices II, LU decomposition  The inverse of a non-singular block matrix can be computed by Ã ! 1 P S R Q 1 ³ 1 ´ 1 ³ P S·´ P SQ R 1 1 = ³ ´ Q RP S Q RP 1S 1 · ³ ´ 1 Q RP 1S RP 1 This is established merely by multiplying the orig- inal block matrix by its alleged inverse, and verify- ing that the product is indeed the identity matrix. The previous identity is used here. This is a very useful result if onP or Q is much smaller than the other, which in turn has special struc- ture which makes it easy to invert.  Example: ³ ´ 1 110 I + 110 = I 1 + 62 so that Ã ! 1 Ã ¡ ¢ ! I 1 I+ 110 1 1 ( + 1) 1 0 1 = 1 0 ( + 1) 1 ( + 1) Ã ! 1 ( + 1) I 11 0 1 = + 1 10 1  Another consequence of (10.1) isA is an × matrix and B is a × matrix ( ), then the eigenvalues oBA are those oAB together with zeros. Proof: Using the two formulas for the determi- nant of a block matrix, we obtain Ã ! I A ¯ 1 ¯ det B I = | I |¯ AB ¯ ¯ ¯ = ¯ 1AB ¯ = |I AB | and = |I || I BA | = |I BA | Thus, if and are the characteristic poly- nomials, we have ( ) ( ) = ( ) 63 and the roots of the characteristic equa( ) = 0 foBA are those of the characteristic equation ( ) = 0 forAB together with 0s - the matriceAB and BA have the same non- zero eigenvalues; i= then alleigenvalues
More Less

Related notes for STAT312
Me

OR

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Join to view

OR

By registering, I agree to the Terms and Privacy Policies
Just a few more details

So we can recommend you notes for your school.