D D 1/2 L (LD 1/2 L ) A = LL T A P AP T = LDL T (P, L, D ) ( [2], [3] ) A 0 A = LDL T (L, D ) ( Cholesky ) A A = LL T (L ) ( Cholesky ) Trefethen and
|
|
- あまめ にばし
- 7 years ago
- Views:
Transcription
1 Cholesky , Cholesky Sylvester L U U 1, Cholesky LU ( Hermite Hermite ) LU LU Cholesky Cholesky A A = LDU (L, D, U ) LDU LDU A = A T = U T D T L T = U T DL T L = U T, U = L T A = LDL T. 1
2 D D 1/2 L (LD 1/2 L ) A = LL T A P AP T = LDL T (P, L, D ) ( [2], [3] ) A 0 A = LDL T (L, D ) ( Cholesky ) A A = LL T (L ) ( Cholesky ) Trefethen and Bau III [9] [7], [5] 2 Cholesky 2.1 A = (a ij ) M(N; R) L = (l ij ) M(N; R) A = LL T A Cholesky (Cholesky factorization) Cholesky (1) Cholesky LU 1 decompose factorization 2
3 2.1 (LU ) N N (1) A GL(N; R) P GL(N; R), L GL(N; R), U GL(N; R) (1) P A = LU (2) a 0 A GL(N; R) L GL(N; R), U GL(N; R) (2) A = LU (3) (1), (2) LU LU = L DU (L, U, D ) ( LDU ) a MATLAB det A(1 : r, 1 : r) A r A = (a ij ) GL(N; R) 0 LDU L = (l ij ) GL(N; R), U = (u ij ) GL(N; R), D = diag (d 1, d 2,, d N ) GL(N; R) A = LDU LDU = A = A T = (LDU) T = U T D T L T = U T DL T (U T, L T ) LDU U = L T. A = LDL T Sylvester D d i (L 1 A(L 1 ) T = D ) ( D 1/2 = diag d1, d 2,, ) d N D ( (D 1/2 ) 2 = D) L = LD 1/2 L L T = (LD 1/2 )(LD 1/2 ) T = LD 1/2 (D 1/2 ) T L T = LD 1/2 D 1/2 L T = LDL T = A. 3
4 2.3 Cholesky A = LL T L D ( ) L 2.4 Cholesky 2.2 Cholesky Cholesky A = LL T (3) a ij = N l ik l jk (i, j = 1, 2,, N) k=1 L = (l ij ) i < j = l ij = 0 (3) a ij = min(i,j) k=1 l ik l jk (i, j = 1, 2,, N) A i j (3) a ij = i a 11 = l 2 11, min(i,j) k=1 a 21 = l 21 l 11, a 22 = l l 2 22, l ik l jk (1 j i N). a 31 = l 31 l 11, a 32 = l 31 l 21 + l 32 l 22, a 33 = l l l 2 33, a i1 = l i1 l 11, a i2 = l i1 l 21 + l i2 l 22,, a ii = l 2 i1 + l 2 i2 + + l 2 ii,. a N1 = l N1 l 11, a N2 = l N1 l 21 + l N2 l 22,, a NN = l 2 N1 + l 2 N2 + + l 2 NN. 4
5 l 11 = ± a 11, l 21 = a 21 /l 11, l 22 = ± a 22 l 2 21, l 31 = a 31 /l 11, l 32 = (a 32 l 31 l 21 ) /l 22, l 33 = ± a 33 l 2 31 l 2 32,. l i1 = a i1 /l 11, l i2 = (a i2 l i1 l 21 ) /a i2, ( ) j 1 l ij = a ij l ik l jk /l jj (j = 1, 2,, i 1),. k=1 l ii = ± a ii l 2 i1 l2 i2 l2 i,i 1, l N1 = a N1 /l 11, l N2 = (a N2 l N1 l 21 ) /a N2, ( ) j 1 l Nj = a Nj l Nk l jk /l jj (j = 1, 2,, N 1), k=1 l NN = ± a NN l 2 N1 l2 N2 l2 N,N 1 ( ). A = (a ij ) (i j ) L = (l ij ) (i j ) ( 0 ) Cholesky (ijk ) for (i = 1; i <= N; i++) { /* L[i][j] (i>j) */ for (j = 1; j < i; j++) { s = a[i][j]; for (k = 1; k < j; k++) s -= L[i][k] * L[j][k]; L[i][j] = s / L[j][j]; } /* L[i][i] */ s = a[i][i]; for (k = 1; k < i; k++) s -= sqr(l[i][k]); L[i][i] = sqrt(s); } l 11, l 21, l 22, l 31,, l 33,, l N1, l N2,, l NN ( ) jik l 11, l 21,, l N1, l 22, l 23,, l N2,, l N 1,N 1, l N,N 1, l NN ( ) ( ) 5
6 Cholesky (jik ) for (j = 1; j <= n; j++) { /* L[j][j] */ s = a[j][j]; for (k = 1; k < j; k++) s -= sqr(l[j][k]); if (s < 0) { fprintf(stderr, "s < 0\n"); exit(0); } L[j][j] = sqrt(s); /* L[i][j] (i>j) */ for (i = j + 1; i <= n; i++) { s = a[i][j]; for (k = 1; k < j; k++) s -= L[i][k] * L[j][k]; L[i][j] = s / L[j][j]; } } } ( [5], Trefethen and Bau III [9], Higham [1] ) A = LL T A = R T R ( R = L T ) i j Cholesky A = R T R R T R = A i j (i, j) r 1i r 1j + r 2i r 2j + + r ii r ij = a ij ( ) i 1 r ii = aii rkj 2, r i 1 ij = a ij r ki r kj /r ii (i > j). k=1 (A a ij ) k=1 2.5 L A = LL T A Cholesky L 6
7 2.5.1 cholesky1.c 1 /* 2 * cholesky1.c 3 */ 4 5 #include <stdio.h> 6 #include <math.h> 7 #include <matrix.h> 8 #include <limits.h> /* INT_MAX */ 9 10 /* [0,1) */ 11 double drandom() 12 { 13 return random() / (double) INT_MAX; 14 } /* */ 17 void mul_mm(int n, matrix ab, matrix a, matrix b) 18 { 19 int i, j, k; 20 double s; 21 for (i = 1; i <= n; i++) 22 for (j = 1; j <= n; j++) { 23 s = 0; 24 for (k = 1; k <= n; k++) 25 s += a[i][k] * b[k][j]; 26 ab[i][j] = s; 27 } 28 } /* */ 31 void print_matrix(int n, matrix a) 32 { 33 int i, j; 34 for (i = 1; i <= n; i++) { 35 for (j = 1; j <= n; j++) 36 printf("%f ", a[i][j]); 37 printf("\n"); 38 } 39 } void clear_m(int n, matrix a) 42 { 43 int i, j; 44 for (i = 1; i <= n; i++) 45 for (j = 1; j <= n; j++) 46 a[i][j] = 0; 47 } /* */ 50 double sqr(double x) { return x * x; } /* Cholesky */ 53 void cholesky(int n, matrix L, matrix a) 54 { 7
8 55 int i, j, k; 56 double s; 57 for (i = 1; i <= n; i++) { 58 for (j = 1; j < i; j++) { 59 s = a[i][j]; 60 for (k = 1; k < j; k++) 61 s -= L[i][k] * L[j][k]; 62 L[i][j] = s / L[j][j]; 63 } 64 /* L[i][i] */ 65 s = a[i][i]; 66 for (k = 1; k < i; k++) 67 s -= sqr(l[i][k]); 68 if (s < 0) { 69 fprintf(stderr, "s < 0\n"); 70 exit(0); 71 } 72 L[i][i] = sqrt(s); 73 } 74 } int main() 77 { 78 int i, j, N; 79 matrix a, L, Lt; 80 printf("n="); scanf("%d", &N); 81 a = new_matrix(n+1, N+1); 82 L = new_matrix(n+1, N+1); 83 Lt = new_matrix(n+1, N+1); 84 for (i = 1; i <= N; i++) { 85 for (j = 1; j <= i; j++) 86 Lt[j][i] = L[i][j] = drandom(); 87 for (j = i + 1; j <= N; j++) 88 Lt[j][i] = L[i][j] = 0.0; 89 } 90 mul_mm(n, a, L, Lt); 91 printf("l=\n"); 92 print_matrix(n, L); 93 printf("a=\n"); 94 print_matrix(n, a); clear_m(n, L); 97 cholesky(n, L, a); 98 printf("l=\n"); 99 print_matrix(n, L); return 0; 102 } 8
9 2.5.2 cholesky1 oyabun%./cholesky1 N=4 L= a= L= oyabun% jik Cholesky 1 /* Cholesky */ 2 void cholesky(int n, matrix L, matrix a) 3 { 4 int i, j, k; 5 double s; 6 for (j = 1; j <= n; j++) { 7 /* L[j][j] */ 8 s = a[j][j]; 9 for (k = 1; k < j; k++) 10 s -= sqr(l[j][k]); 11 if (s < 0) { 12 fprintf(stderr, "s < 0\n"); 13 exit(0); 14 } 15 L[j][j] = sqrt(s); 16 for (i = j + 1; i <= n; i++) { 17 s = a[i][j]; 18 for (k = 1; k < j; k++) 19 s -= L[i][k] * L[j][k]; 20 L[i][j] = s / L[j][j]; 21 } 22 } 23 } 9
10 2.5.4 jik oyabun%./cholesky2 N=4 L= a= L= oyabun% 2.6 Cholesky U = L T L L[i][j] u[j][i] A ( ) 1 /* Cholesky A=U^T U (A ) */ 2 void cholesky(int n, matrix U, matrix a) 3 { 4 int j, i, k; 5 double s; 6 for (i = 1; i <= n; i++) { 7 /* U[i][i] */ 8 s = a[i][i]; 9 for (k = 1; k < i; k++) 10 s -= sqr(u[k][i]); 11 if (s < 0) { 12 fprintf(stderr, "s < 0\n"); 13 exit(0); 14 } 15 U[i][i] = sqrt(s); 16 for (j = i + 1; j <= n; j++) { 17 s = a[i][j]; 18 for (k = 1; k < i; k++) 19 s -= U[k][j] * U[k][i]; 20 U[i][j] = s / U[i][i]; 21 } 22 } 23 } 10
11 2.7 Cholesky (2) Trefethen and Bau III [9] ( ) A (1, 1) 1 Gauss 1 ( ) ( ) ( ) 1 w T w T A = = w K w I 0 K ww T ( 1 w T 0 K ww T ) = ( K ww T ) ( 1 w T 0 I ) A = ( 1 0 w I ) ( A = K ww T ( a 11 α = a 11 ( ) ( α T A = w/α I 0 K ww T /a 11 w w T B ) ) ( ) ( 1 w T 0 I α 0 w/α I ) ). = R T 1 A 1 R 1 A 1 = R T 2 A 2 R 2, A 2 = R T 3 A 3 R 3,, A N 1 = R T NA N R N A N = I (N ) A = R T 1 R T 2 R T NR N R 2 R 1. R = R N R 2 R 1 Cholesky R = A for k = 1 to m for j=k+1 to m R j,j:m = R j,j:m R k,j:m R kj /R kk R k,k:m = R k,k:m / R kk A = R T R 11
12 2.8 N 3 /6 + O(N 2 ), N 3 /3 + O(N 2 ) Gauss N O(N) LU L U 1 Cholesky 1 N (O(N 2 ) N ) Cholesky 2.9 Trefethen and Bau III [9] ( ) Cholesky 2.7 Cholesky Gauss Cholesky ( 2.2 ) R 2 R 2 = R T 2 = A 1/2 ( ) p (1 p ) N 1 N A 1/2 R = R T N A 1/2. (Cf. LU partial pivoting ( L = 1 ) U = 2 N 1 A ) 2.2 (Cholesky ) A N (A), (B) Cholesky ε M R (4) δa M(N; R) s.t. RT R = A + δa, δa A = O(ε M) (ε M 0). 2.3 O( ) Higham [1] p.197 Theorem 10.3 δa γ N+1 R T R γ n γ n = nu (u the unit round off) 1 nu IEEE 754 u = ε M 12
13 R R R R = O(κ(A)ε M ) R 2.10 Ax = b 1 LU A Ax = b A A = LL T Cholesky LL T x = b y Ly = b L T x = y x O(N 2 ) ( ) 2.4 (Cholesky 1 ) 1 Ax = b (A), (B) Cholesky x A M(N; R) s.t. (A + A) x = b, A A = O(ε M) 2.5 Higham [1] p.198, Theorem 10.4 (O( ) ) A γ 3N+1 R T R A 2 γ 3N+1 N(1 Nγ N+1 ) 1 A 2 ( 1 ) 13
14 3 Cholesky ( ) 3.1 ( Cholesky ) N A = (a ij ) δ k (k = 1, 2,, N) 0 A = LDL T (L, D ) 1 d 1 l L = l 31 l 32 1 d 2, D = l N1 l N2 l N,N 1 0 d N 1 LDL T (i, j) (l i,1 l i,2 l i,i ) d 1 l j1 d 2 l j2. d j 1 l j,j 1 d j 0. 0 = min{i,j} k=1 d k l ik l jk ( l ii = 1). A i j l ii = 1 i 1 i d i + d k l 2 ik (j = i) k=1 a ij = d k l ik l jk = i 1 k=1 d k l ik l jk + d i l ji (j = i + 1, i + 2,, N) k=1 i = 1, 2,, N i 1 d i = a ii d k l 2 ik, k=1 ) l ji = 1 i 1 (a ij d k l ik l jk d i k=1 14 (j = i + 1,, N)
15 {d i }, {l ji } i j d i Gauss i d (i) ii 0 d 1 = a 11 = δ 1 0, l j1 = a j1 /d 1 (j = 2,, N), d 2 = a 22 d 1 l 2 21 = δ 2 /δ 1 0, l j2 = (a 2j d 1 l 21 l j1 )/d 2 (j = 3,, N), d 3 = a 33 d 1 l 2 31 d 2 l 2 32 = δ 3 /δ 2 0, l j3 = (a 3j d 1 l 31 l j1 d 2 l 32 l j2 )/d 3 (j = 4,, N), A Cholesky LU ( [6] ) 4 ( ) Cholesky LU A Cholesky A = L T L 1 Ax = b Ly = b, L T x = y 15
16 1 /* */ 2 /* Ux=b */ 3 void solve_uxb(int n, matrix u, vector b) 4 { 5 int i, j; 6 for (i = n; i >= 1; i--) { 7 for (j = i + 1; j <= n; j++) 8 b[i] -= u[i][j] * b[j]; 9 b[i] /= u[i][i]; 10 } 11 } /* L^T x=b solve_uxb() u[i][j] L[j][i] */ 14 void solve_ltxb(int n, matrix L, vector b) 15 { 16 int i, j; 17 for (i = n; i >= 1; i--) { 18 for (j = i + 1; j <= n; j++) 19 b[i] -= L[j][i] * b[j]; 20 b[i] /= L[i][i]; 21 } 22 } /* Lx=b */ 25 void solve_lxb(int n, matrix L, vector b) 26 { 27 int i, j; 28 for (i = 1; i <= n; i++) { 29 for (j = 1; j < i; j++) 30 b[i] -= L[i][j] * b[j]; 31 b[i] /= L[i][i]; 32 } 33 } /* U^T x=b */ 36 void solve_utxb(int n, matrix U, vector b) 37 { 38 int i, j; 39 for (i = 1; i <= n; i++) { 40 for (j = 1; j < i; j++) 41 b[i] -= U[j][i] * b[j]; 42 b[i] /= U[i][i]; 43 } 44 } 16
17 oyabun%./cholesky3 N=5 L= a= L= x= oyabun% A Cholesky A.1 C 1 /* 2 * band_cholesky.c --- A Cholesky 3 */ 4 5 #include <stdio.h> 6 #include <limits.h> 7 #include <math.h> 8 #include <matrix.h> 9 10 double drandom() { return random() / (double) INT_MAX; } 11 double sqr(double x) { return x * x; } 12 double max(double a, double b) { return (a > b)? a : b; } 13 double min(double a, double b) { return (a < b)? a : b; } /* Cholesky A=U^T U (A, U ) 16 * 17 * m n A=(a_{ij}) (1 i n, 1 j n) 18 * ASB=(asb_{ij) (1 i n, 1 j m+1) 19 * 20 * 21 * asb[i][j-i+1] = a[i][j] (1 i n, i j min(i+m,n)) 22 * 23 * U ASB 24 * 17
18 25 */ 26 void band_cholesky(int n, int m, matrix ASB) 27 { 28 int j, i, k; 29 double s; 30 for (i = 1; i <= n; i++) { 31 /* U[i][i] */ 32 s = ASB[i][1]; 33 for (k = max(1,i-m); k < i; k++) 34 s -= sqr(asb[k][i-k+1]); 35 if (s < 0) { 36 fprintf(stderr, "s < 0\n"); 37 exit(0); 38 } 39 ASB[i][1] = sqrt(s); 40 /* U[i][j] (i < j) */ 41 for (j = i + 1; j <= min(i + m, n); j++) { 42 s = ASB[i][j-i+1]; 43 for (k = max(1,j-m); k < i; k++) 44 s -= ASB[k][j-k+1] * ASB[k][i-k+1]; 45 ASB[i][j-i+1] = s / ASB[i][1]; 46 } 47 } 48 } /* */ 51 /* U^T x=b */ 52 void solve_band_utxb(int n, int m, matrix UB, vector b) 53 { 54 int i, j; 55 for (i = 1; i <= n; i++) { 56 for (j = max(i-m,1); j < i; j++) 57 b[i] -= UB[j][i-j+1] * b[j]; 58 b[i] /= UB[i][1]; 59 } 60 } /* Ux=b */ 63 void solve_band_uxb(int n, int m, matrix UB, vector b) 64 { 65 int i, j; 66 for (i = n; i >= 1; i--) { 67 for (j = i + 1; j <= min(i+m,n); j++) 68 b[i] -= UB[i][j-i+1] * b[j]; 69 b[i] /= UB[i][1]; 70 } 71 } /* */ 74 void print_symmetric_band_matrix1(int n, int m, matrix asb) 75 { 76 int i, j; 77 for (i = 1; i <= n; i++) { 78 for (j = 1; j <= n; j++) 79 if (abs(j - i) > m) 80 printf("%f ", 0.0); 18
19 81 else { 82 if (j >= i) 83 printf("%f ", asb[i][j-i+1]); 84 else /* i j */ 85 printf("%f ", asb[j][i-j+1]); 86 } 87 printf("\n"); 88 } 89 } /* */ 92 void print_upper_triangular_band_matrix1(int n, int m, matrix ub) 93 { 94 int i, j; 95 for (i = 1; i <= n; i++) { 96 for (j = 1; j <= n; j++) 97 if (abs(j - i) > m) 98 printf("%f ", 0.0); 99 else { 100 if (j >= i) 101 printf("%f ", ub[i][j-i+1]); 102 else 103 printf("%f ", 0.0); 104 } 105 printf("\n"); 106 } 107 } void print_vector1(int n, vector x) 110 { 111 int i; 112 for (i = 1; i <= n; i++) 113 printf("%f ", x[i]); 114 printf("\n"); 115 } int main() 118 { 119 int n, m, i, j, k; 120 double s; 121 matrix ASB, UB; 122 vector x, b; /*, */ 125 n = 8; m = 3; /* ASB, UB n (m+1) */ 128 ASB = new_matrix(n + 1, (m+1) + 1); 129 UB = new_matrix(n + 1, (m+1) + 1); 130 x = new_vector(n + 1); 131 b = new_vector(n + 1); /* U */ 134 for (i = 1; i <= n; i++) 135 for (j = i; j <= min(i+m, n); j++) 136 UB[i][j-i+1] = drandom(); 19
20 137 printf("u=\n"); 138 print_upper_triangular_band_matrix1(n, m, UB); /* A=U^T U */ 141 for (i = 1; i <= n; i++) 142 for (j = i; j <= min(i+m,n); j++) { /* */ 143 s = 0; 144 /* U[k][i]*U[k][j] 145 max(i-m,1) k i max(j-m,1) k j i j max(j-m,1) k i 147 */ 148 for (k = max(j-m,1); k <= i; k++) 149 s += UB[k][i-k+1] * UB[k][j-k+1]; 150 ASB[i][j-i+1] = s; 151 } 152 printf("a:=u^t U\n"); 153 print_symmetric_band_matrix1(n, m, ASB); /* x */ 156 for (i = 1; i <= n; i++) 157 x[i] = i; /* b=a x */ 160 for (i = 1; i <= n; i++) { 161 s = ASB[i][1] * x[i]; 162 for (j = i + 1; j <= min(i+m,n); j++) 163 s += ASB[i][j-i+1] * x[j]; 164 for (j = max(i-m,1); j < i; j++) 165 s += ASB[j][i-j+1] * x[j]; 166 b[i] = s; 167 } /* Cholesky A=U^T U */ 170 band_cholesky(n, m, ASB); 171 printf("cholesky U=\n"); 172 print_upper_triangular_band_matrix1(n, m, ASB); /* U^T U x = b */ 175 solve_band_utxb(n, m, UB, b); 176 solve_band_uxb(n, m, UB, b); 177 printf("x=\n"); 178 print_vector1(n, b); return 0; 181 } 20
21 oyabun%./band_cholesky U= A:=U^T U Cholesky U= x= oyabun% A.2 C++ & Profil Profil 1 /* 2 * band_cholesky_c++.c --- A Cholesky 3 */ 4 5 #include <iostream.h> 6 #include <iomanip.h> 7 #include <limits.h> 8 #include <math.h> 9 #include <Matrix.h> 10 #include <Vector.h> double drandom(void) { return random() / (double) INT_MAX; } 13 double sqr(double x) { return x * x; } 14 // double max(double a, double b) { return (a > b)? a : b; } 15 // double min(double a, double b) { return (a < b)? a : b; } /* Cholesky A=U^T U (A, U ) 21
22 18 * 19 * m n A=(a_{ij}) (1 i n, 1 j n) 20 * ASB=(asb_{ij) (1 i n, 1 j m+1) 21 * 22 * 23 * asb[i][j-i+1] = a[i][j] (1 i n, i j min(i+m,n)) 24 * 25 * U ASB 26 * 27 */ 28 void band_cholesky(int n, int m, MATRIX &ASB) 29 { 30 double s; 31 for (int i = 1; i <= n; i++) { 32 /* U[i][i] */ 33 s = ASB(i,1); 34 for (int k = max(1,i-m); k < i; k++) 35 s -= sqr(asb(k,i-k+1)); 36 if (s < 0) { 37 cerr << "s < 0" << endl; 38 exit(0); 39 } 40 ASB(i,1) = sqrt(s); 41 /* U[i][j] (i < j) */ 42 for (int j = i + 1; j <= min(i + m, n); j++) { 43 s = ASB(i,j-i+1); 44 for (int k = max(1,j-m); k < i; k++) 45 s -= ASB(k,j-k+1) * ASB(k,i-k+1); 46 ASB(i,j-i+1) = s / ASB(i,1); 47 } 48 } 49 } /* */ 52 /* U^T x=b */ 53 void solve_band_utxb(int n, int m, const MATRIX &UB, VECTOR &b) 54 { 55 for (int i = 1; i <= n; i++) { 56 for (int j = max(i-m,1); j < i; j++) 57 b(i) -= UB(j,i-j+1) * b(j); 58 b(i) /= UB(i,1); 59 } 60 } /* Ux=b */ 63 void solve_band_uxb(int n, int m, const MATRIX &UB, VECTOR &b) 64 { 65 for (int i = n; i >= 1; i--) { 66 for (int j = i + 1; j <= min(i+m,n); j++) 67 b(i) -= UB(i,j-i+1) * b(j); 68 b(i) /= UB(i,1); 69 } 70 } /* */ 73 void print_symmetric_band_matrix1(int n, int m, const MATRIX &asb) 22
23 74 { 75 for (int i = 1; i <= n; i++) { 76 for (int j = 1; j <= n; j++) 77 if (abs(j - i) > m) 78 cout << 0.0 << " "; 79 else { 80 if (j >= i) 81 cout << asb(i,j-i+1) << " "; 82 else /* i j */ 83 cout << asb(j,i-j+1) << " "; 84 } 85 cout << endl; 86 } 87 } /* */ 90 void print_upper_triangular_band_matrix1(int n, int m, const MATRIX &ub) 91 { 92 for (int i = 1; i <= n; i++) { 93 for (int j = 1; j <= n; j++) 94 if (abs(j - i) > m) 95 cout << 0.0 << " "; 96 else { 97 if (j >= i) 98 cout << ub(i,j-i+1) << " "; 99 else 100 cout << 0.0 << " "; 101 } 102 cout << endl; 103 } 104 } void print_vector1(int n, const VECTOR &x) 107 { 108 for (int i = 1; i <= n; i++) 109 cout << x(i) << " "; 110 cout << endl; 111 } int main() 114 { 115 int n, m, i, j, k; 116 double s; cout << setprecision(4); 119 cout << setiosflags(ios::fixed); 120 /*, */ 121 n = 8; m = 3; /* ASB, UB n (m+1) */ 124 MATRIX ASB(n,m+1), UB(n,m+1); 125 VECTOR x(n), b(n); /* U */ 128 for (i = 1; i <= n; i++) 129 for (j = i; j <= min(i+m, n); j++) 23
24 130 UB(i,j-i+1) = drandom(); 131 cout << "U=" << endl; 132 print_upper_triangular_band_matrix1(n, m, UB); /* A=U^T U */ 135 for (i = 1; i <= n; i++) 136 for (j = i; j <= min(i+m,n); j++) { /* */ 137 s = 0; 138 /* U[k](i)*U[k](j) 139 max(i-m,1) k i max(j-m,1) k j i j max(j-m,1) k i 141 */ 142 for (k = max(j-m,1); k <= i; k++) 143 s += UB(k,i-k+1) * UB(k,j-k+1); 144 ASB(i,j-i+1) = s; 145 } 146 cout << "A:=U^T U" << endl; 147 print_symmetric_band_matrix1(n, m, ASB); /* x */ 150 for (i = 1; i <= n; i++) 151 x(i) = i; /* b=a x */ 154 for (i = 1; i <= n; i++) { 155 s = ASB(i,1) * x(i); 156 for (j = i + 1; j <= min(i+m,n); j++) 157 s += ASB(i,j-i+1) * x(j); 158 for (j = max(i-m,1); j < i; j++) 159 s += ASB(j,i-j+1) * x(j); 160 b(i) = s; 161 } /* Cholesky A=U^T U */ 164 band_cholesky(n, m, ASB); 165 cout << "Cholesky U=" << endl; 166 print_upper_triangular_band_matrix1(n, m, ASB); /* U^T U x = b */ 169 solve_band_utxb(n, m, UB, b); 170 solve_band_uxb(n, m, UB, b); 171 cout << "x=" << endl; 172 print_vector1(n, b); return 0; 175 } 24
25 oyabun% g++ -W -I/usr/local/include -o band_cholesky_c++ band_cholesky_c++.c -lprofil -lmatrix oyabun%./band_cholesky_c++ U= A:=U^T U Cholesky U= x= oyabun% B Trefethen and Bau III [9] ( ) F (A) F fl: R F (B) x R ε R s.t. ε ε M, fl(x) = x(1 + ε). F,,, {+,,, }, x, y F ε R s.t. ε ε M, x y = (x y)(1 + ε). 25
26 C 2 C.1 R N R N (, ) R N (, ) x = (x i ) R N, y = (y i ) R N (x, y) R N = (x, y) = N x i y i. i=1 y y T x : (x, y) R N = y T x. C.2 R N a: R N R N R 1. x R N, y R N, z R N a(x + y, z) = a(x, z) + a(y, z), a(x, y + z) = a(x, z) + a(y, z). 2. x R N, y R N, λ R a(λx, y) = λa(x, y), a(x, λy) = λa(x, y). N A = (a ij ) M(N; R) a(x, y) def. = (Ax, y) R N R N R N a a ij = a(e j, e i ) ( N ) N a x j e j, y i e i = j=1 i=1 A = (a ij ) N N x j y i a(e j, e i ) = j=1 i=1 ( N N ) a ij x j y i = (Ax, y) R N. i=1 j=1 C.3 R N x R N y R N a(x, y) = a(y, x) a R N a(x, y) = (Ax, y) R N a(x, y) = (Ax, y) R N = y T Ax = (A T y) T x = (x, A T y) R N = (A T y, x) R N, 26
27 a(y, x) = (Ay, x) R N a x R N y R N (A T y, x) R N = (Ay, x) R N A = A T. R N A (Ax, y) C.4 R N R N a(, ) x R N a(x, x) 0 ( x R N a(x, x) = 0 x = 0 ). a(, ) R N a A a(x, y) = (Ax, y) R N C.5 2 a: R N R R N 2 2 A = (a ij ) N N a(x) = a ij x i x j, i=1 j=1 a(x) = (Ax, x) R N (Ax, x) R N = (x, A T x) R N = (A T x, x) R N A A T (Ax, x) R N = 1 ( ) 1 2 ((Ax, x) R N + (AT x, x) R N ) = 2 (A + AT )x, x A (A + A T )/2 ( 2 x (Ax, x) A ) (Ax, x) = 1 2 (A + AT )x = Ax R N 27
28 R N 2 R N (x, y) (x, x) a(x, y) = 1 2 (a(x + y, x + y) a(x, x) a(y, y)) = 1 (a(x + y) a(x) a(y)) 2 2 C.6 A = (a ij ) M(N; R) 2 f(x) = (Ax, x) = x T Ax P GL(N; R), x = P y f(x) = f(p y) = (P y) T A(P y) = y T P T AP y = (P T AP y, y) = (A y, y). A := P T AP 2 (A ) A diag (λ 1,, λ N ) f(x) = N λ i yi 2. i=1 λ i A C.1 (Sylvester (Sylvester s law of inertia)) A n P D := P T AP n D A D Sylvester ( 2013/8/26 10:52:43) D.1 Sylvester 2 ( ) 28
29 2 (2 ) A P P T AP A P T AP A P 1 AP ( ) (P T = P 1 ) ( ) 1,2 2 Taylor 2 2 ( [4]) f n a f (a) = 0 f(a + h) = f(a) + h 2 + O( h 2 ) (h 0). 2 Morse Lemma 2 Sylvester D.2 3 ( ) A A 0 π(a), ν(a), ζ(a) A n π(a), ν(a), ζ(a) 0, π(a) + ν(a) + ζ(a) = n ( ) ( Sylvester ) A n U (U 1 = U T ) λ 1,, λ n U T AU = diag (λ 1,, λ n ). x R n y := U 1 x x = Uy (Ax, x) = (AUy, Uy) = (U T AUy, y) = n λ j yj 2. 2 (Ax, x) ( ) j=1
30 ( ) (i) U T AU diag (λ 1,, λ n ) (ii) U (Ax, x) U ( U P ) U λ j A λ j (j = 1,, n) A Sylvester D.1 (Sylvester, (Sylvester s law of inertia)) A P P T AP = diag (λ 1,, λ n ) n p := {j {1,, n}; λ j > 0}, n n := {j {1,, n}; λ j < 0}, n z := {j {1,, n}; λ j = 0} ( ) P A ( 0 A ) D.2 ( Sylvester ) A P P T AP = diag (λ 1,, λ n ) n p := {j {1,, n}; λ j > 0}, n n := {j {1,, n}; λ j < 0}, n z := {j {1,, n}; λ j = 0} n p = π(a), n n = ν(a), n z = ζ(a). 30
31 D.1 P U T AU = diag[λ 1,, λ n ] U P T AP = U T AU = U 1 AU A λ 1,, λ n A n p, n n, n z π(a), ν(a), ζ(a) [8] Sylvester D.3 ( Sylvester ) A P B := P T AP π(a) = π(b), ν(a) = ν(b), ζ(a) = ζ(b) D.2 B ( ) D.2 D.3 ( ) D.3 n A, P λ 1,, λ n s.t. P T AP = diag[λ 1,, λ n ] B := P T AP ( ) λ 1,, λ n n p = π(b) = π(a), n n = ν(b) = ν(a), n z = ζ(b) = ζ(a). D.2 A, P B := P T AP B U : λ 1,, λ n R s.t. U T BU = diag[λ 1,, λ n ]. diag (λ 1,, λ n ) = U T BU = U T P T AP U = (P U) T A(P U) = P T AP, P := P U. P U P D.2 (B ) λ 1,, λ n 0 π(a), ν(a), ζ(a) π(b) = π(a), ν(b) = ν(a), ζ(b) = ζ(a). D.3 Sylvester D.1 D.1 ( ) x = Sy, x = Rz (Ax, x) = α 1 y α r y 2 r α r+1 y 2 r+1 α p y 2 p (α j > 0) = β 1 z β s z 2 s β s+1 z 2 s+1 β q z 2 q (β j > 0) 31
32 p q p = q, r = s y j, z k x y j z k 1 r < s y 1,, y p y 1,, y r, z s+1,, z q 1 ( p 1 y 1,, y p r + (q s) 1 r + (q s) = (r s) + q < q p ) y k, r + 1 k p y 1,, y r, z s+1,, z q y 1 = = y r = z s+1 = = z q = 0, y k = 1 x 1,, x n 1 (Ax, x) = α α r 0 2 α r+1 y 2 r+1 α k 1 2 α p y 2 p α k < 0, (Ax, x) = β 1 z β s z 2 s β s β q r > s y 1,, y r y r+1,, y p, z 1,, z s y k (1 k r) y r+1,, y p, z 1,, z s y r+1 = = y p = z 1 = = z s = 0, y k = 1 x 1,, x n 1 (Ax, x) = α 1 y α k α r y 2 r α r α p 0 2 α k > 0, (Ax, x) = β β s 0 2 β s+1 y 2 s+1 β q β 2 q 0 r = s. ( Ax, x) ( ) D.2 ( ) D.2 A n P := {V ; V R n, x V \ {0} (Ax, x) > 0}, N := {V ; V R n, x V \ {0} (Ax, x) < 0}, Z := {V ; V R n, x V \ {0} (Ax, x) = 0} {0} P, N, Z P, N, Z =. N p := max{dim V ; V P}, N n := max{dim V ; V N }, N z := max{dim V ; V Z} ( : dim{0} = 0 ) : N p = π(a), N n = ν(a), N z = ζ(a). 32
33 A v 1,, v n Av j = λ j v j V p := Span{v j ; λ j > 0}, V n := Span{v j ; λ j < 0}, V z := Span{v j ; λ j = 0} V p P, V n N, V z Z. dim V p = π(a), dim V n = ν(a), dim V z = ζ(a) N p, N n, N z ( ) N p π(a), N n ν(a), N z ζ(a). W p P, W n N, W z Z W p W n = W N W z = W z W p = {0} W p, W n, W z ( ) dim W p = N p, dim W n = N n, dim W z = N z W p W n W z R n n = π(a) + ν(a) + ζ(a) N p + N n + N z n. N p = π(a), N n = ν(a), N z = ζ(a) ( ). P x = P y (Ax, x) = n µ j yj 2 p q 0, (Ax, x) = α 1 y α p y 2 p α p+1 y 2 p+1 α p+q y 2 p+q, α j > 0 (1 j p + q) (a) p {P y; y p+1 = = y n = 0} (Ax, x) > 0 p π(a). (b) q {P y; y 1 = = y p = y p+q+1 = = y n = 0} (Ax, x) < 0 q ν(a). (c) n (p+q) {Sy; y 1 = = y p+q = 0} (Ax, x) = 0 n (p+q) ζ(a). (c) p + q + ζ(a) n (a) π(a) p, (b) ν(a) q j=1 n = π(a) + ν(a) + ζ(a) p + q + ζ(a) n. p = π(a), q = ν(a). 33
34 [1] Nicholas J. Higham. Accuracy and Stability of Numerical Algorithms. SIAM, [2] [3].., 2003., I, II,, (1993, 1994)..., [4]. 1 (2013 ). jp/~mk/lecture/tahensuu1-2013/tahensuu1-new-text.pdf, [5].., [6].. 7., [7],,..., [8],.., [9] Lloyd Nicholas Trefethen and David Bau III. Numerical Linear Algebra. SIAM,
Gauss
15 1 LU LDL T 6 : 1g00p013-5 1 6 1.1....................................... 7 1.2.................................. 8 1.3.................................. 8 2 Gauss 9 2.1.....................................
More information1 5 13 4 1 41 1 411 1 412 2 413 3 414 3 415 4 42 6 43 LU 7 431 LU 10 432 11 433 LU 11 44 12 441 13 442 13 443 SOR ( ) 14 444 14 445 15 446 16 447 SOR 16 448 16 45 17 4 41 n x 1,, x n a 11 x 1 + a 1n x
More informationGauss Strassen LU LU LU LU 22 5 Gauss LU
I, 208 3 23 3 2 4 3 6 3 6 32 6 32 6 322 7 323 Gauss 8 324 Strassen 9 325 0 326 0 327 LU 0 33 4 LU 2 4 2 4 2 42 3 43 4 42 8 43 0 LU 20 44 LU 22 5 Gauss LU 23 5 23 52 Gauss 23 53 LU 24 53 24 532 25 533 LU
More informationA11 (1993,1994) 29 A12 (1994) 29 A13 Trefethen and Bau Numerical Linear Algebra (1997) 29 A14 (1999) 30 A15 (2003) 30 A16 (2004) 30 A17 (2007) 30 A18
2013 8 29y, 2016 10 29 1 2 2 Jordan 3 21 3 3 Jordan (1) 3 31 Jordan 4 32 Jordan 4 33 Jordan 6 34 Jordan 8 35 9 4 Jordan (2) 10 41 x 11 42 x 12 43 16 44 19 441 19 442 20 443 25 45 25 5 Jordan 26 A 26 A1
More informationLINEAR ALGEBRA I Hiroshi SUZUKI Department of Mathematics International Christian University
LINEAR ALGEBRA I Hiroshi SUZUKI Department of Mathematics International Christian University 2002 2 2 2 2 22 2 3 3 3 3 3 4 4 5 5 6 6 7 7 8 8 9 Cramer 9 0 0 E-mail:hsuzuki@icuacjp 0 3x + y + 2z 4 x + y
More information行列代数2010A
(,) A (,) B C = AB a 11 a 1 a 1 b 11 b 1 b 1 c 11 c 1 c a A = 1 a a, B = b 1 b b, C = AB = c 1 c c a 1 a a b 1 b b c 1 c c i j ij a i1 a i a i b 1j b j b j c ij = a ik b kj b 1j b j AB = a i1 a i a ik
More informationII 2 3.,, A(B + C) = AB + AC, (A + B)C = AC + BC. 4. m m A, m m B,, m m B, AB = BA, A,, I. 5. m m A, m n B, AB = B, A I E, 4 4 I, J, K
II. () 7 F 7 = { 0,, 2, 3, 4, 5, 6 }., F 7 a, b F 7, a b, F 7,. (a) a, b,,. (b) 7., 4 5 = 20 = 2 7 + 6, 4 5 = 6 F 7., F 7,., 0 a F 7, ab = F 7 b F 7. (2) 7, 6 F 6 = { 0,, 2, 3, 4, 5 },,., F 6., 0 0 a F
More informationexample2_time.eps
Google (20/08/2 ) ( ) Random Walk & Google Page Rank Agora on Aug. 20 / 67 Introduction ( ) Random Walk & Google Page Rank Agora on Aug. 20 2 / 67 Introduction Google ( ) Random Walk & Google Page Rank
More information2 (2016 3Q N) c = o (11) Ax = b A x = c A n I n n n 2n (A I n ) (I n X) A A X A n A A A (1) (2) c 0 c (3) c A A i j n 1 ( 1) i+j A (i, j) A (i, j) ã i
[ ] (2016 3Q N) a 11 a 1n m n A A = a m1 a mn A a 1 A A = a n (1) A (a i a j, i j ) (2) A (a i ca i, c 0, i ) (3) A (a i a i + ca j, j i, i ) A 1 A 11 0 A 12 0 0 A 1k 0 1 A 22 0 0 A 2k 0 1 0 A 3k 1 A rk
More information0.,,., m Euclid m m. 2.., M., M R 2 ψ. ψ,, R 2 M.,, (x 1 (),, x m ()) R m. 2 M, R f. M (x 1,, x m ), f (x 1,, x m ) f(x 1,, x m ). f ( ). x i : M R.,,
2012 10 13 1,,,.,,.,.,,. 2?.,,. 1,, 1. (θ, φ), θ, φ (0, π),, (0, 2π). 1 0.,,., m Euclid m m. 2.., M., M R 2 ψ. ψ,, R 2 M.,, (x 1 (),, x m ()) R m. 2 M, R f. M (x 1,, x m ), f (x 1,, x m ) f(x 1,, x m ).
More information数学Ⅱ演習(足助・09夏)
II I 9/4/4 9/4/2 z C z z z z, z 2 z, w C zw z w 3 z, w C z + w z + w 4 t R t C t t t t t z z z 2 z C re z z + z z z, im z 2 2 3 z C e z + z + 2 z2 + 3! z3 + z!, I 4 x R e x cos x + sin x 2 z, w C e z+w
More information2000 7 17 iii,,.,,,,.,.,,. =,,.,,.,.,,,,,,, fortran,.,,,,.,,, iv (i),., 10,, 10 1. (ii),, fortran, fortran, C,.. (i),., 10 23.. (ii),.,, fortran Pasal,, for, if then else. C., fortran C.,.,,.,.,,.,,.,.,,,,.
More information(Basic Theory of Information Processing) 1
(Basic Theory of Information Processing) 1 10 (p.178) Java a[0] = 1; 1 a[4] = 7; i = 2; j = 8; a[i] = j; b[0][0] = 1; 2 b[2][3] = 10; b[i][j] = a[2] * 3; x = a[2]; a[2] = b[i][3] * x; 2 public class Array0
More information1 Ricci V, V i, W f : V W f f(v ) = Imf W ( ) f : V 1 V k W 1
1 Ricci V, V i, W f : V W f f(v = Imf W ( f : V 1 V k W 1 {f(v 1,, v k v i V i } W < Imf > < > f W V, V i, W f : U V L(U; V f : V 1 V r W L(V 1,, V r ; W L(V 1,, V r ; W (f + g(v 1,, v r = f(v 1,, v r
More informationPart. 4. () 4.. () 4.. 3 5. 5 5.. 5 5.. 6 5.3. 7 Part 3. 8 6. 8 6.. 8 6.. 8 7. 8 7.. 8 7.. 3 8. 3 9., 34 9.. 34 9.. 37 9.3. 39. 4.. 4.. 43. 46.. 46..
Cotets 6 6 : 6 6 6 6 6 6 7 7 7 Part. 8. 8.. 8.. 9..... 3. 3 3.. 3 3.. 7 3.3. 8 Part. 4. () 4.. () 4.. 3 5. 5 5.. 5 5.. 6 5.3. 7 Part 3. 8 6. 8 6.. 8 6.. 8 7. 8 7.. 8 7.. 3 8. 3 9., 34 9.. 34 9.. 37 9.3.
More information弾性定数の対称性について
() by T. oyama () ij C ij = () () C, C, C () ij ji ij ijlk ij ij () C C C C C C * C C C C C * * C C C C = * * * C C C * * * * C C * * * * * C () * P (,, ) P (,, ) lij = () P (,, ) P(,, ) (,, ) P (, 00,
More information(I) GotoBALS, http://www-is.amp.i.kyoto-u.ac.jp/ kkimur/charpoly.html 2
sdmp Maple - (Ver.2) ( ) September 27, 2011 1 (I) GotoBALS, http://www-is.amp.i.kyoto-u.ac.jp/ kkimur/charpoly.html 2 (II) Nehalem CPU GotoBLAS Intel CPU Nehalem CPU, GotoBLAS, Hyper-Thread technology
More information8.1 Fubini 8.2 Fubini 9 (0%) 10 (50%) 10.1 10.2 Carathéodory 10.3 Fubini 1 Introduction [1],, [2],, [3],, [4],, [5],, [6],, [7],, [8],, [1, 2, 3] 1980
% 100% 1 Introduction 2 (100%) 2.1 2.2 2.3 3 (100%) 3.1 3.2 σ- 4 (100%) 4.1 4.2 5 (100%) 5.1 5.2 5.3 6 (100%) 7 (40%) 8 Fubini (90%) 2006.11.20 1 8.1 Fubini 8.2 Fubini 9 (0%) 10 (50%) 10.1 10.2 Carathéodory
More information(2016 2Q H) [ ] R 2 2 P = (a, b), Q = (c, d) Q P QP = ( ) a c b d (a c, b d) P = (a, b) O P ( ) a p = b P = (a, b) p = ( ) a b R 2 {( ) } R 2 x = x, y
(2016 2Q H) [ ] R 2 2 P = (a, b), Q = (c, d) Q P QP = a c b d (a c, b d) P = (a, b) O P a p = b P = (a, b) p = a b R 2 { } R 2 x = x, y R y 2 a p =, c q = b d p + a + c q = b + d q p P q a p = c R c b
More information●70974_100_AC009160_KAPヘ<3099>ーシス自動車約款(11.10).indb
" # $ % & ' ( ) * +, -. / 0 1 2 3 4 5 6 7 8 9 : ; < = >? @ A B C D E F G H I J K L M N O P Q R S T U V W X Y " # $ % & ' ( ) * + , -. / 0 1 2 3 4 5 6 7 8 9 : ; < = > ? @ A B
More information(2018 2Q C) [ ] R 2 2 P = (a, b), Q = (c, d) Q P QP = ( ) a c b d (a c, b d) P = (a, b) O P ( ) a p = b P = (a, b) p = ( ) a b R 2 {( ) } R 2 x = x, y
(2018 2Q C) [ ] R 2 2 P = (a, b), Q = (c, d) Q P QP = a c b d (a c, b d) P = (a, b) O P a p = b P = (a, b) p = a b R 2 { } R 2 x = x, y R y 2 a p =, c q = b d p + a + c q = b + d q p P q a p = c R c b
More information,..,,.,,.,.,..,,.,,..,,,. 2
A.A. (1906) (1907). 2008.7.4 1.,.,.,,.,,,.,..,,,.,,.,, R.J.,.,.,,,..,.,. 1 ,..,,.,,.,.,..,,.,,..,,,. 2 1, 2, 2., 1,,,.,, 2, n, n 2 (, n 2 0 ).,,.,, n ( 2, ), 2 n.,,,,.,,,,..,,. 3 x 1, x 2,..., x n,...,,
More information3 3 i
00D8102021I 2004 3 3 3 i 1 ------------------------------------------------------------------------------------------------1 2 ---------------------------------------------------------------------------------------2
More informationC言語による数値計算プログラミング演習
5. 行列の固有値問題 n n 正方行列 A に対する n 個の固有値 λ i (i=1,,,n) と対応する固有ベクトル u i は次式を満たす Au = λ u i i i a11 a1 L a1 n u1i a1 a a n u i A =, ui = M O M M an 1 an L ann uni これらはまとめて, つぎのように書ける 5.1 ヤコビ法 = Λ, = [ u1 u u
More information1 n A a 11 a 1n A =.. a m1 a mn Ax = λx (1) x n λ (eigenvalue problem) x = 0 ( x 0 ) λ A ( ) λ Ax = λx x Ax = λx y T A = λy T x Ax = λx cx ( 1) 1.1 Th
1 n A a 11 a 1n A = a m1 a mn Ax = λx (1) x n λ (eigenvalue problem) x = ( x ) λ A ( ) λ Ax = λx x Ax = λx y T A = λy T x Ax = λx cx ( 1) 11 Th9-1 Ax = λx λe n A = λ a 11 a 12 a 1n a 21 λ a 22 a n1 a n2
More information離散最適化基礎論 第 11回 組合せ最適化と半正定値計画法
11 okamotoy@uec.ac.jp 2019 1 25 2019 1 25 10:59 ( ) (11) 2019 1 25 1 / 38 1 (10/5) 2 (1) (10/12) 3 (2) (10/19) 4 (3) (10/26) (11/2) 5 (1) (11/9) 6 (11/16) 7 (11/23) (11/30) (12/7) ( ) (11) 2019 1 25 2
More information24 I ( ) 1. R 3 (i) C : x 2 + y 2 1 = 0 (ii) C : y = ± 1 x 2 ( 1 x 1) (iii) C : x = cos t, y = sin t (0 t 2π) 1.1. γ : [a, b] R n ; t γ(t) = (x
24 I 1.1.. ( ) 1. R 3 (i) C : x 2 + y 2 1 = 0 (ii) C : y = ± 1 x 2 ( 1 x 1) (iii) C : x = cos t, y = sin t (0 t 2π) 1.1. γ : [a, b] R n ; t γ(t) = (x 1 (t), x 2 (t),, x n (t)) ( ) ( ), γ : (i) x 1 (t),
More informationDecember 28, 2018
e-mail : kigami@i.kyoto-u.ac.jp December 28, 28 Contents 2............................. 3.2......................... 7.3..................... 9.4................ 4.5............. 2.6.... 22 2 36 2..........................
More information1 4 1 ( ) ( ) ( ) ( ) () 1 4 2
7 1995, 2017 7 21 1 2 2 3 3 4 4 6 (1).................................... 6 (2)..................................... 6 (3) t................. 9 5 11 (1)......................................... 11 (2)
More information5 36 5................................................... 36 5................................................... 36 5.3..............................
9 8 3............................................. 3.......................................... 4.3............................................ 4 5 3 6 3..................................................
More information第85 回日本感染症学会総会学術集会後抄録(III)
β β α α α µ µ µ µ α α α α γ αβ α γ α α γ α γ µ µ β β β β β β β β β µ β α µ µ µ β β µ µ µ µ µ µ γ γ γ γ γ γ µ α β γ β β µ µ µ µ µ β β µ β β µ α β β µ µµ β µ µ µ µ µ µ λ µ µ β µ µ µ µ µ µ µ µ
More informationd ϕ i) t d )t0 d ϕi) ϕ i) t x j t d ) ϕ t0 t α dx j d ) ϕ i) t dx t0 j x j d ϕ i) ) t x j dx t0 j f i x j ξ j dx i + ξ i x j dx j f i ξ i x j dx j d )
23 M R M ϕ : R M M ϕt, x) ϕ t x) ϕ s ϕ t ϕ s+t, ϕ 0 id M M ϕ t M ξ ξ ϕ t d ϕ tx) ξϕ t x)) U, x 1,...,x n )) ϕ t x) ϕ 1) t x),...,ϕ n) t x)), ξx) ξ i x) d ϕi) t x) ξ i ϕ t x)) M f ϕ t f)x) f ϕ t )x) fϕ
More informationX G P G (X) G BG [X, BG] S 2 2 2 S 2 2 S 2 = { (x 1, x 2, x 3 ) R 3 x 2 1 + x 2 2 + x 2 3 = 1 } R 3 S 2 S 2 v x S 2 x x v(x) T x S 2 T x S 2 S 2 x T x S 2 = { ξ R 3 x ξ } R 3 T x S 2 S 2 x x T x S 2
More informationPart () () Γ Part ,
Contents a 6 6 6 6 6 6 6 7 7. 8.. 8.. 8.3. 8 Part. 9. 9.. 9.. 3. 3.. 3.. 3 4. 5 4.. 5 4.. 9 4.3. 3 Part. 6 5. () 6 5.. () 7 5.. 9 5.3. Γ 3 6. 3 6.. 3 6.. 3 6.3. 33 Part 3. 34 7. 34 7.. 34 7.. 34 8. 35
More informationver Web
ver201723 Web 1 4 11 4 12 5 13 7 2 9 21 9 22 10 23 10 24 11 3 13 31 n 13 32 15 33 21 34 25 35 (1) 27 4 30 41 30 42 32 43 36 44 (2) 38 45 45 46 45 5 46 51 46 52 48 53 49 54 51 55 54 56 58 57 (3) 61 2 3
More informationv v = v 1 v 2 v 3 (1) R = (R ij ) (2) R (R 1 ) ij = R ji (3) 3 R ij R ik = δ jk (4) i=1 δ ij Kronecker δ ij = { 1 (i = j) 0 (i
1. 1 1.1 1.1.1 1.1.1.1 v v = v 1 v 2 v 3 (1) R = (R ij ) (2) R (R 1 ) ij = R ji (3) R ij R ik = δ jk (4) δ ij Kronecker δ ij = { 1 (i = j) 0 (i j) (5) 1 1.1. v1.1 2011/04/10 1. 1 2 v i = R ij v j (6) [
More information+ 1 ( ) I IA i i i 1 n m a 11 a 1j a 1m A = a i1 a ij a im a n1 a nj a nm.....
+ http://krishnathphysaitama-uacjp/joe/matrix/matrixpdf 1 ( ) I IA i i i 1 n m a 11 a 1j a 1m A = a i1 a ij a im a n1 a nj a nm (1) n m () (n, m) ( ) n m B = ( ) 3 2 4 1 (2) 2 2 ( ) (2, 2) ( ) C = ( 46
More informationQR
1 7 16 13 1 13.1 QR...................................... 2 13.1.1............................................ 2 13.1.2..................................... 3 13.1.3 QR........................................
More informationC による数値計算法入門 ( 第 2 版 ) 新装版 サンプルページ この本の定価 判型などは, 以下の URL からご覧いただけます. このサンプルページの内容は, 新装版 1 刷発行時のものです.
C による数値計算法入門 ( 第 2 版 ) 新装版 サンプルページ この本の定価 判型などは, 以下の URL からご覧いただけます. http://www.morikita.co.jp/books/mid/009383 このサンプルページの内容は, 新装版 1 刷発行時のものです. i 2 22 2 13 ( ) 2 (1) ANSI (2) 2 (3) Web http://www.morikita.co.jp/books/mid/009383
More informationJanuary 27, 2015
e-mail : kigami@i.kyoto-u.ac.jp January 27, 205 Contents 2........................ 2.2....................... 3.3....................... 6.4......................... 2 6 2........................... 6
More information1 8, : 8.1 1, 2 z = ax + by + c ax by + z c = a b +1 x y z c = 0, (0, 0, c), n = ( a, b, 1). f = n i=1 a ii x 2 i + i<j 2a ij x i x j = ( x, A x), f =
1 8, : 8.1 1, z = ax + by + c ax by + z c = a b +1 x y z c = 0, (0, 0, c), n = ( a, b, 1). f = a ii x i + i
More information20 9 19 1 3 11 1 3 111 3 112 1 4 12 6 121 6 122 7 13 7 131 8 132 10 133 10 134 12 14 13 141 13 142 13 143 15 144 16 145 17 15 19 151 1 19 152 20 2 21 21 21 211 21 212 1 23 213 1 23 214 25 215 31 22 33
More informationC , C++ C C++ C++ C cpprefjp - C++ 1 C CUI 2.1 donothing.cpp 1
C++ 2018 7 1, 2018 11 4 http://nalab.mind.meiji.ac.jp/~mk/labo/text/nantoka-c++/ 1 C++ C C++ C++ C cpprefjp - C++ 1 C++17 2 2 CUI 2.1 donothing.cpp 1 /* 2 * donothing.cpp 3 */ 4 5 int main() 6 7 return
More informationuntitled
1 1 Ax = b A R m m A b R m x R m A shift-and invert Lanczos - LU CG A = LU LU Ly = b Ux = y A LU A A = LL T 1 LU b,, Vol. 11, No. 4, pp. 14 18 (2006). x * x (0), x (1), x (2), A Ap A # x (n+1) = Cx (n)
More informationI , : ~/math/functional-analysis/functional-analysis-1.tex
I 1 2004 8 16, 2017 4 30 1 : ~/math/functional-analysis/functional-analysis-1.tex 1 3 1.1................................... 3 1.2................................... 3 1.3.....................................
More informationnumb.dvi
11 Poisson kanenko@mbkniftycom alexeikanenko@docomonejp http://wwwkanenkocom/ , u = f, ( u = u+f u t, u = f t ) 1 D R 2 L 2 (D) := {f(x,y) f(x,y) 2 dxdy < )} D D f,g L 2 (D) (f,g) := f(x,y)g(x,y)dxdy (L
More informationx, y x 3 y xy 3 x 2 y + xy 2 x 3 + y 3 = x 3 y xy 3 x 2 y + xy 2 x 3 + y 3 = 15 xy (x y) (x + y) xy (x y) (x y) ( x 2 + xy + y 2) = 15 (x y)
x, y x 3 y xy 3 x 2 y + xy 2 x 3 + y 3 = 15 1 1977 x 3 y xy 3 x 2 y + xy 2 x 3 + y 3 = 15 xy (x y) (x + y) xy (x y) (x y) ( x 2 + xy + y 2) = 15 (x y) ( x 2 y + xy 2 x 2 2xy y 2) = 15 (x y) (x + y) (xy
More informationi I II I II II IC IIC I II ii 5 8 5 3 7 8 iii I 3........................... 5......................... 7........................... 4........................ 8.3......................... 33.4...................
More information( )/2 hara/lectures/lectures-j.html 2, {H} {T } S = {H, T } {(H, H), (H, T )} {(H, T ), (T, T )} {(H, H), (T, T )} {1
( )/2 http://www2.math.kyushu-u.ac.jp/ hara/lectures/lectures-j.html 1 2011 ( )/2 2 2011 4 1 2 1.1 1 2 1 2 3 4 5 1.1.1 sample space S S = {H, T } H T T H S = {(H, H), (H, T ), (T, H), (T, T )} (T, H) S
More informationlinearal1.dvi
19 4 30 I 1 1 11 1 12 2 13 3 131 3 132 4 133 5 134 6 14 7 2 9 21 9 211 9 212 10 213 13 214 14 22 15 221 15 222 16 223 17 224 20 3 21 31 21 32 21 33 22 34 23 341 23 342 24 343 27 344 29 35 31 351 31 352
More informationad bc A A A = ad bc ( d ) b c a n A n A n A A det A A ( ) a b A = c d det A = ad bc σ {,,,, n} {,,, } {,,, } {,,, } ( ) σ = σ() = σ() = n sign σ sign(
I n n A AX = I, YA = I () n XY A () X = IX = (YA)X = Y(AX) = YI = Y X Y () XY A A AB AB BA (AB)(B A ) = A(BB )A = AA = I (BA)(A B ) = B(AA )B = BB = I (AB) = B A (BA) = A B A B A = B = 5 5 A B AB BA A
More information13 0 1 1 4 11 4 12 5 13 6 2 10 21 10 22 14 3 20 31 20 32 25 33 28 4 31 41 32 42 34 43 38 5 41 51 41 52 43 53 54 6 57 61 57 62 60 70 0 Gauss a, b, c x, y f(x, y) = ax 2 + bxy + cy 2 = x y a b/2 b/2 c x
More informationテクノ東京21 2003年6月号(Vol.123)
2 3 5 7 9 10 11 12 13 - 21 2003 6123 21 2003 6123 - 21 2003 6123 21 2003 6123 3 u x y x Ax Bu y Cx Du uy x A,B,C,D - 21 2003 6123 21 2003 6123 - 21 2003 6123 - 21 2003 6123 -- -- - 21 2003 6123 03 3832-3655
More informationオートマトンと言語理論 テキスト 成蹊大学理工学部情報科学科 山本真基 ii iii 1 1 1.1.................................. 1 1.2................................ 5 1.3............................. 5 2 7 2.1..................................
More informationii
ii iii 1 1 1.1..................................... 1 1.2................................... 3 1.3........................... 4 2 9 2.1.................................. 9 2.2...............................
More informationx V x x V x, x V x = x + = x +(x+x )=(x +x)+x = +x = x x = x x = x =x =(+)x =x +x = x +x x = x ( )x = x =x =(+( ))x =x +( )x = x +( )x ( )x = x x x R
V (I) () (4) (II) () (4) V K vector space V vector K scalor K C K R (I) x, y V x + y V () (x + y)+z = x +(y + z) (2) x + y = y + x (3) V x V x + = x (4) x V x + x = x V x x (II) x V, α K αx V () (α + β)x
More information行列代数2010A
a ij i j 1) i +j i, j) ij ij 1 j a i1 a ij a i a 1 a j a ij 1) i +j 1,j 1,j +1 a i1,1 a i1,j 1 a i1,j +1 a i1, a i +1,1 a i +1.j 1 a i +1,j +1 a i +1, a 1 a,j 1 a,j +1 a, ij i j 1,j 1,j +1 ij 1) i +j a
More informationuntitled
0. =. =. (999). 3(983). (980). (985). (966). 3. := :=. A A. A A. := := 4 5 A B A B A B. A = B A B A B B A. A B A B, A B, B. AP { A, P } = { : A, P } = { A P }. A = {0, }, A, {0, }, {0}, {}, A {0}, {}.
More information2S III IV K A4 12:00-13:30 Cafe David 1 2 TA 1 appointment Cafe David K2-2S04-00 : C
2S III IV K200 : April 16, 2004 Version : 1.1 TA M2 TA 1 10 2 n 1 ɛ-δ 5 15 20 20 45 K2-2S04-00 : C 2S III IV K200 60 60 74 75 89 90 1 email 3 4 30 A4 12:00-13:30 Cafe David 1 2 TA 1 email appointment Cafe
More information一般演題(ポスター)
6 5 13 : 00 14 : 00 A μ 13 : 00 14 : 00 A β β β 13 : 00 14 : 00 A 13 : 00 14 : 00 A 13 : 00 14 : 00 A β 13 : 00 14 : 00 A β 13 : 00 14 : 00 A 13 : 00 14 : 00 A β 13 : 00 14 : 00 A 13 : 00 14 : 00 A
More information7 9 7..................................... 9 7................................ 3 7.3...................................... 3 A A. ω ν = ω/π E = hω. E
B 8.9.4, : : MIT I,II A.P. E.F.,, 993 I,,, 999, 7 I,II, 95 A A........................... A........................... 3.3 A.............................. 4.4....................................... 5 6..............................
More information2 2 MATHEMATICS.PDF 200-2-0 3 2 (p n ), ( ) 7 3 4 6 5 20 6 GL 2 (Z) SL 2 (Z) 27 7 29 8 SL 2 (Z) 35 9 2 40 0 2 46 48 2 2 5 3 2 2 58 4 2 6 5 2 65 6 2 67 7 2 69 2 , a 0 + a + a 2 +... b b 2 b 3 () + b n a
More information漸化式のすべてのパターンを解説しましたー高校数学の達人・河見賢司のサイト
https://www.hmg-gen.com/tuusin.html https://www.hmg-gen.com/tuusin1.html 1 2 OK 3 4 {a n } (1) a 1 = 1, a n+1 a n = 2 (2) a 1 = 3, a n+1 a n = 2n a n a n+1 a n = ( ) a n+1 a n = ( ) a n+1 a n {a n } 1,
More information1 Introduction 1 (1) (2) (3) () {f n (x)} n=1 [a, b] K > 0 n, x f n (x) K < ( ) x [a, b] lim f n (x) f(x) (1) f(x)? (2) () f(x)? b lim a f n (x)dx = b
1 Introduction 2 2.1 2.2 2.3 3 3.1 3.2 σ- 4 4.1 4.2 5 5.1 5.2 5.3 6 7 8. Fubini,,. 1 1 Introduction 1 (1) (2) (3) () {f n (x)} n=1 [a, b] K > 0 n, x f n (x) K < ( ) x [a, b] lim f n (x) f(x) (1) f(x)?
More information受賞講演要旨2012cs3
アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート アハ ート α β α α α α α
More information2008 ( 13 ) C LAPACK 2008 ( 13 )C LAPACK p. 1
2008 ( 13 ) C LAPACK LAPACK p. 1 Q & A Euler http://phase.hpcc.jp/phase/mppack/long.pdf KNOPPIX MT (Mersenne Twister) SFMT., ( ) ( ) ( ) ( ). LAPACK p. 2 C C, main Asir ( Asir ) ( ) (,,...), LAPACK p.
More information1 Abstract 2 3 n a ax 2 + bx + c = 0 (a 0) (1) ( x + b ) 2 = b2 4ac 2a 4a 2 D = b 2 4ac > 0 (1) 2 D = 0 D < 0 x + b 2a = ± b2 4ac 2a b ± b 2
1 Abstract n 1 1.1 a ax + bx + c = 0 (a 0) (1) ( x + b ) = b 4ac a 4a D = b 4ac > 0 (1) D = 0 D < 0 x + b a = ± b 4ac a b ± b 4ac a b a b ± 4ac b i a D (1) ax + bx + c D 0 () () (015 8 1 ) 1. D = b 4ac
More informationII K116 : January 14, ,. A = (a ij ) ij m n. ( ). B m n, C n l. A = max{ a ij }. ij A + B A + B, AC n A C (1) 1. m n (A k ) k=1,... m n A, A k k
: January 14, 28..,. A = (a ij ) ij m n. ( ). B m n, C n l. A = max{ a ij }. ij A + B A + B, AC n A C (1) 1. m n (A k ) k=1,... m n A, A k k, A. lim k A k = A. A k = (a (k) ij ) ij, A k = (a ij ) ij, i,
More informationuntitled
3,,, 2 3.1 3.1.1,, A4 1mm 10 1, 21.06cm, 21.06cm?, 10 1,,,, i),, ),, ),, x best ± δx 1) ii), x best ), δx, e,, e =1.602176462 ± 0.000000063) 10 19 [C] 2) i) ii), 1) x best δx
More information2012 A, N, Z, Q, R, C
2012 A, N, Z, Q, R, C 1 2009 9 2 2011 2 3 2012 9 1 2 2 5 3 11 4 16 5 22 6 25 7 29 8 32 1 1 1.1 3 1 1 1 1 1 1? 3 3 3 3 3 3 3 1 1, 1 1 + 1 1 1+1 2 2 1 2+1 3 2 N 1.2 N (i) 2 a b a 1 b a < b a b b a a b (ii)
More information応用数学III-4.ppt
III f x ( ) = 1 f x ( ) = P( X = x) = f ( x) = P( X = x) =! x ( ) b! a, X! U a,b f ( x) =! " e #!x, X! Ex (!) n! ( n! x)!x! " x 1! " x! e"!, X! Po! ( ) n! x, X! B( n;" ) ( ) ! xf ( x) = = n n!! ( n
More informationhttp://www2.math.kyushu-u.ac.jp/~hara/lectures/lectures-j.html 2 N(ε 1 ) N(ε 2 ) ε 1 ε 2 α ε ε 2 1 n N(ɛ) N ɛ ɛ- (1.1.3) n > N(ɛ) a n α < ɛ n N(ɛ) a n
http://www2.math.kyushu-u.ac.jp/~hara/lectures/lectures-j.html 1 1 1.1 ɛ-n 1 ɛ-n lim n a n = α n a n α 2 lim a n = 1 n a k n n k=1 1.1.7 ɛ-n 1.1.1 a n α a n n α lim n a n = α ɛ N(ɛ) n > N(ɛ) a n α < ɛ
More information20 6 4 1 4 1.1 1.................................... 4 1.1.1.................................... 4 1.1.2 1................................ 5 1.2................................... 7 1.2.1....................................
More information: : : : ) ) 1. d ij f i e i x i v j m a ij m f ij n x i =
1 1980 1) 1 2 3 19721960 1965 2) 1999 1 69 1980 1972: 55 1999: 179 2041999: 210 211 1999: 211 3 2003 1987 92 97 3) 1960 1965 1970 1985 1990 1995 4) 1. d ij f i e i x i v j m a ij m f ij n x i = n d ij
More information1 4 2 EP) (EP) (EP)
2003 2004 2 27 1 1 4 2 EP) 5 3 6 3.1.............................. 6 3.2.............................. 6 3.3 (EP)............... 7 4 8 4.1 (EP).................... 8 4.1.1.................... 18 5 (EP)
More information24.15章.微分方程式
m d y dt = F m d y = mg dt V y = dy dt d y dt = d dy dt dt = dv y dt dv y dt = g dv y dt = g dt dt dv y = g dt V y ( t) = gt + C V y ( ) = V y ( ) = C = V y t ( ) = gt V y ( t) = dy dt = gt dy = g t dt
More informationTabulation of the clasp number of prime knots with up to 10 crossings
. Tabulation of the clasp number of prime knots with up to 10 crossings... Kengo Kawamura (Osaka City University) joint work with Teruhisa Kadokami (East China Normal University).. VI December 20, 2013
More informationA
A 2563 15 4 21 1 3 1.1................................................ 3 1.2............................................. 3 2 3 2.1......................................... 3 2.2............................................
More informationsim98-8.dvi
8 12 12.1 12.2 @u @t = @2 u (1) @x 2 u(x; 0) = (x) u(0;t)=u(1;t)=0fort 0 1x, 1t N1x =1 x j = j1x, t n = n1t u(x j ;t n ) Uj n U n+1 j 1t 0 U n j =1t=(1x) 2 = U n j+1 0 2U n j + U n j01 (1x) 2 (2) U n+1
More informationall.dvi
29 4 Green-Lagrange,,.,,,,,,.,,,,,,,,,, E, σ, ε σ = Eε,,.. 4.1? l, l 1 (l 1 l) ε ε = l 1 l l (4.1) F l l 1 F 30 4 Green-Lagrange Δz Δδ γ = Δδ (4.2) Δz π/2 φ γ = π 2 φ (4.3) γ tan γ γ,sin γ γ ( π ) γ tan
More information8.1 Fubini 8.2 Fubini 9 (0%) 10 (50%) Carathéodory 10.3 Fubini 1 Introduction 1 (1) (2) {f n (x)} n=1 [a, b] K > 0 n, x f n (x) K < ( ) x [a
% 100% 1 Introduction 2 (100%) 2.1 2.2 2.3 3 (100%) 3.1 3.2 σ- 4 (100%) 4.1 4.2 5 (100%) 5.1 5.2 5.3 6 (100%) 7 (40%) 8 Fubini (90%) 2007.11.5 1 8.1 Fubini 8.2 Fubini 9 (0%) 10 (50%) 10.1 10.2 Carathéodory
More information( ) a C n ( R n ) R a R C n. a C n (or R n ) a 0 2. α C( R ) a C n αa = α a 3. a, b C n a + b a + b ( ) p 8..2 (p ) a = [a a n ] T C n p n a
9 8 m n mn N.J.Nigham, Accuracy and Stability of Numerical Algorithms 2nd ed., (SIAM) x x = x2 + y 2 = x + y = max( x, y ) x y x () (norm) (condition number) 8. R C a, b C a b 0 a, b a = a 0 0 0 n C n
More information