Introduction to Linear algebra-Strang-Solutions-Manual_ver13

A tridiagonal matrix a has bidiagonal factors l and u

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 0 1 1 140 41 0 4 4D4 1 4 0 1 1 D LDLT . 011 004 011 4 001 2 32 32 3 aaaa 1 a a a a a ¤ 0 All of the b a b a b a7 b ¤ a multipliers 6a b b b 7 61 1 76 . Need 4 a b c c 5 D 4 1 1 1 54 c b c b5 c ¤ b are `ij D 1 abcd 1111 dc d ¤ c for this A 2 32 32 3 arrr 1 a r r r a¤0 b r s r s r7 b¤r 6a b s s 7 61 1 76 . Need 4a b c t 5 D 41 1 1 54 c s t s5 c¤s abcd 1111 dt d ¤t         5 2 24 2 2 10 . gives x D xD . Then gives c D cD 3 3 01 3 11 41       2 24 2 24 D c. xD . Forward to xD Ax D b is LU x D 3 01 11 8 17 "# # "# " "# # "# " 3 4 111 4 4 100 1 1 0 c D 5 gives c D 1 . Then 0 1 1 x D 1 gives x D 0 . 1 1 001 1 6 111 Those are the forward elimination and back substitution steps for # "# #" " 4 111 1 1 1 xD 5 . Ax D 1 1 6 1 111 (b) I goes to L 1 (c) LU goes to U . Elimination multiply by L 1 ! 18 (a) Multiply LDU D L1 D1 U1 by inverses to get L1 1 LD D D1 U1 U 1 . The left side is lower triangular, the right side is upper triangular ) both sides are diagonal. (b) L; U; L1 ; U1 have diagonal 1’s so D D D1 . Then L1 1 L and U1 U 1 are both I . # " # # " " #" a 1 110 a a 0 b b 1 1 D LI U I a a C b D (same L) 19 1 1 0 b bCc c 1 011 (same U ). A tridiagonal matrix A has bidiagonal factors L and U . 20 A tridiagonal T has 2 nonzeros in the pivot row and only one nonzero below the pivot (one operation to find ` and then one for the new pivot!). T D bidiagonal L times bidiagonal U . 17 (a) L goes to I Solutions to Exercises 23 21 For the first matrix A; L keeps the 3 lower zeros at the start of rows. But U may not 22 23 24 25 26 have the upper zero where A24 D 0. For the second matrix B; L keeps the bottom left zero at the start of row 4. U keeps the upper right zero at the start of column 4. One zero in A and two zeros in B are filled in. " # " # " # 531 420 200 Eliminating upwards, 3 3 1 ! 2 2 0 ! 2 2 0 D L. We reach 111 111 111 a lower triangular L, and the multipliers are in an upper triangular U . A D UL with " # 111 UD 0 1 1 . 001 The 2 by 2 upper submatrix A2 has the first two pivots 5; 9. Reason: Elimination on A starts in the upper left corner with elimination on A2 . The upper left blocks all factor at the same time as A: Ak is Lk Uk . The i; j entry of L 1 is j= i for i  j . And Li i 1 is .1 i /= i below the diagonal .K 1 /ij D j.n i C 1/=.n C 1/ for i  j (and symmetric): .n C 1/K 1 looks good. Problem Set 2.7, page 115  1 1 AD 9  1 AD c      0 19 1 0 T 1 has A D ;A D ; .A 1 /T D .AT / 3 03 3 1=3    10 c c has AT D A and A 1 D 2 D .A 1 /T . 0 1 cc 1   1 3 D ; 0 1=3 2 .AB/T is not AT B T except when AB D BA. Transpose that to find: B T AT D AT B T . 1T / D .B 1 A 1 /T D .A 1 /T .B 1 /T . This is also .AT / 1 .B T / 1 . (b) If U is upper triangular, so is U 1 : then .U 1 /T is lower triangular.   01 AD has A2 D 0. The diagonal of AT A has dot products of columns of A with 00 themselves. If AT A D 0, zero dot products ) zero columns ) A D zero matrix.    "0#  123...
View Full Document

This note was uploaded on 09/25/2012 for the course PHY 103 taught by Professor Minki during the Spring '12 term at Korea Advanced Institute of Science and Technology.

Ask a homework question - tutors are online