Macierz odwrotna

Autor podstrony: Krzysztof Zajączkowski

Stronę tą wyświetlono już: 19843 razy

Macierz odwrotna A-1 macierzy A to taka macierz, której iloczyn AA-1 jest równy macierzy jednostkowej I. Wyznaczenie macierzy odwrotnej macierzy A jest możliwe jedynie wtedy, gdy wyznacznik owej macierzy jest różny od zera. Tego typu macierze nazywa się mianem nieosobliwych, zaś macierze, których wyznacznik jest równy zeru nazywa się osobliwymi.

Macierz odwrotną można obliczyć tworząc macierz transponowaną dopełnień algebraicznych AD macierzy A, która podzielona przez wyznacznik macierzy A daje w wyniku macierz odwrotną zgodnie z wzorem [2].

Macierz transponowaną dopełnień algebraicznych AD wyznacza się w następujący sposób:

gdzie:

Przykład

Wyznaczyć macierz odwrotną macierzy A.

Dane:

A=egin{bmatrix}
3 & 2 & 1 & 7 
8 & 2 & 2 & 4 
5 & 0 & 3 & 3 
1 & 0 & 0 & 9
end{bmatrix}

Rozwiązanie:

Przed rozpoczęciem właściwych obliczeń macierzy odwrotnej A-1 oraz macierzy dopełnień algebraicznych AD warto obliczyć wyznacznik macierzy A w celu stwierdzenia czy macierz A jest macierzą nieosobliwą

|A|=egin{vmatrix}
3 & 2 & 1 & 7 
8 & 2 & 2 & 4 
5 & 0 & 3 & 3 
1 & 0 & 0 & 9
end{vmatrix}=-2cdotegin{vmatrix}
8 & 2 & 4 
5 & 3 & 3 
1 & 0 & 9
end{vmatrix}+2cdotegin{vmatrix}
3 & 1 & 7 
5 & 3 & 3 
1 & 0 & 9
end{vmatrix}=-2cdotleft(8cdot 3cdot 9+1cdot 2cdot3-4cdot 3cdot 1-9cdot 2cdot 5
ight)+2cdotleft(3cdot 3cdot 9+1cdot 1cdot 3-7cdot 3cdot 1-9cdot 1cdot 5
ight)=-204

Wyznacznik macierzy A0, więc można przystąpić do wyliczenia argumentów macierzy transponowanej dopełnień algebraicznych AD macierzy A.

Argumenty pierwszego wiersza macierzy AD:

A_{1,1}=(-1)^{1+1}cdot egin{vmatrix}
2 & 2 & 4 
0 & 3 & 3 
0 & 0 & 9
end{vmatrix}=2cdot 3cdot 9=54

A_{1,2}=(-1)^{2+1}cdot egin{vmatrix}
2 & 1 & 7 
0 & 3 & 3 
0 & 0 & 9
end{vmatrix}=-2cdot 3cdot 9=-54

A_{1,3}=(-1)^{3+1}cdot egin{vmatrix}
2 & 1 & 7 
2 & 2 & 4 
0 & 0 & 9
end{vmatrix}=2cdot 2cdot 9-9cdot 1cdot 2=18

A_{1,4}=(-1)^{4+1}cdot egin{vmatrix}
2 & 1 & 7 
2 & 2 & 4 
0 & 3 & 3
end{vmatrix}=-left(2cdot 2cdot 3+2cdot 3cdot 7-4cdot 3cdot 2-3cdot 1cdot 2
ight)=-24

Argumenty drugiego wiersza macierzy AD:

A_{2,1}=(-1)^{1+2}cdot egin{vmatrix}
8 & 2 & 4 
5 & 3 & 3 
1 & 0 & 9
end{vmatrix}=-left(8cdot 3cdot 9+1cdot 2cdot 3-4cdot 3cdot 1-9cdot 2cdot 5
ight)=-120

A_{2,2}=(-1)^{2+2}cdot egin{vmatrix}
3 & 1 & 7 
5 & 3 & 3 
1 & 0 & 9
end{vmatrix}=3cdot 3cdot 9+1cdot 1cdot 3-7cdot 3cdot 1-9cdot 1cdot 5=18

A_{2,3}=(-1)^{3+2}cdot egin{vmatrix}
3 & 1 & 7 
8 & 2 & 4 
1 & 0 & 9
end{vmatrix}=-left(3cdot 2cdot 9+1cdot 1cdot 4-7cdot 2cdot 1-9cdot 1cdot 8
ight)=28

A_{2,4}=(-1)^{4+2}cdot egin{vmatrix}
3 & 1 & 7 
8 & 2 & 4 
5 & 3 & 3
end{vmatrix}=3cdot 2cdot 3+8cdot 3cdot 7+5cdot 1cdot 4-7cdot 2cdot 5-4cdot 3cdot 3-3cdot 1cdot 8=76

Argumenty trzeciego wiersza macierzy AD:

A_{3,1}=(-1)^{1+3}cdot egin{vmatrix}
8 & 2 & 4 
5 & 0 & 3 
1 & 0 & 9
end{vmatrix}=1cdot 2cdot 3-9cdot 2cdot 5=-84

A_{3,2}=(-1)^{2+3}cdot egin{vmatrix}
3 & 2 & 7 
5 & 0 & 3 
1 & 0 & 9
end{vmatrix}=-left(1cdot 2cdot 3-9cdot 2cdot 5
ight)=84

A_{3,3}=(-1)^{3+3}cdot egin{vmatrix}
3 & 2 & 7 
8 & 2 & 4 
1 & 0 & 9
end{vmatrix}=3cdot 2cdot 9+1cdot 2cdot 4-7cdot 2cdot 1-9cdot 2cdot 8=-96

A_{3,4}=(-1)^{4+3}cdot egin{vmatrix}
3 & 2 & 7 
8 & 2 & 4 
5 & 0 & 3
end{vmatrix}=-left(3cdot 2cdot 3+5cdot 2cdot 4-7cdot 2cdot 5-3cdot 2cdot 8
ight)=60

Argumenty czwartego wiersza macierzy AD:

A_{4,1}=(-1)^{1+4}cdot egin{vmatrix}
8 & 2 & 2 
5 & 0 & 3 
1 & 0 & 0
end{vmatrix}=-2cdot 3cdot 1=-6

A_{4,2}=(-1)^{2+4}cdot egin{vmatrix}
3 & 2 & 1 
5 & 0 & 3 
1 & 0 & 0
end{vmatrix}=2cdot 3cdot 1=6

A_{4,3}=(-1)^{3+4}cdot egin{vmatrix}
3 & 2 & 1 
8 & 2 & 2 
1 & 0 & 0
end{vmatrix}=-left(2cdot 2cdot 1-1cdot 2cdot 1
ight)=-2

A_{4,4}=(-1)^{4+4}cdot egin{vmatrix}
3 & 2 & 1 
8 & 2 & 2 
5 & 0 & 3
end{vmatrix}=3cdot 2cdot 3+5cdot 2cdot 2-1cdot 2cdot 5-3cdot 2cdot 8=-20

Ostatecznie macierz transponowana dopełnień algebraicznych AD macierzy A jest równa:

A^D=egin{bmatrix}
A_{1,1} & A_{2,1} & A_{3,1} & A_{4,1} 
A_{1,2} & A_{2,2} & A_{3,2} & A_{4,2} 
A_{1,3} & A_{2,3} & A_{3,3} & A_{4,3} 
A_{1,1} & A_{2,4} & A_{3,4} & A_{4,4}
end{bmatrix}=egin{bmatrix}
54 & -54 & 18 & -24 
-120 & 18 & 28 & 76 
-84 & 84 & -96 & 60 
-6 & 6 & -2 &-20
end{bmatrix}

Macierz odwrotna macierzy A jest więc równa (zgodnie z wzorem [2]):

A^{-1}=frac{A^D}{|A|}=egin{bmatrix}
54 & -54 & 18 & -24 
-120 & 18 & 28 & 76 
-84 & 84 & -96 & 60 
-6 & 6 & -2 &-20
end{bmatrix}cdot left(-frac{1}{204}
ight)=egin{bmatrix}
-cfrac{9}{34} & cfrac{9}{34} & -cfrac{3}{34} & cfrac{2}{17} 
cfrac{10}{17} & -cfrac{3}{34} & -cfrac{7}{51} & -cfrac{19}{51} 
cfrac{7}{17} & -cfrac{7}{17} & cfrac{8}{17} & -cfrac{5}{17} 
cfrac{1}{34} & -cfrac{1}{34} & cfrac{1}{102} &-cfrac{5}{51}
end{bmatrix}

Sprawdzenie

Acdot A^{-1}=egin{bmatrix}
3 & 2 & 1 & 7 
8 & 2 & 2 & 4 
5 & 0 & 3 & 3 
1 & 0 & 0 & 9
end{bmatrix}cdotegin{bmatrix}
-cfrac{9}{34} & cfrac{9}{34} & -cfrac{3}{34} & cfrac{2}{17} 
cfrac{10}{17} & -cfrac{3}{34} & -cfrac{7}{51} & -cfrac{19}{51} 
cfrac{7}{17} & -cfrac{7}{17} & cfrac{8}{17} & -cfrac{5}{17} 
cfrac{1}{34} & -cfrac{1}{34} & cfrac{1}{102} &-cfrac{5}{51}
end{bmatrix}=egin{bmatrix}
1 & 0 & 0 & 0 
0 & 1 & 0 & 0 
0 & 0 & 1 & 0 
0 & 0 & 0 & 1
end{bmatrix}

Korzystając z macierzy odwrotnej można rozwiązać dowolny liniowy układ równań. Dla przykładu stworzę takowy układ na bazie argumentów macierzy A z wcześniejszego przykładu:

egin{cases}
3cdot x_1 + 2cdot x_2 + x_3 + 7cdot x_4=3 
8cdot x_1 + 2cdot x_2 + 2cdot x_3 + 4cdot x_4=4 
5cdot x_1 + 3cdot x_3 + 3cdot x_4=2 
x_1 + 9cdot x_4=0
end{cases}

Mnożąc macierz odwrotną A-1 macierzy A przez macierz wyrazów wolnych uzyskuje się wartości niewiadomych x1, x2, x3, x4 w następujący sposób:

egin{bmatrix}
x_1 
x_2 
x_3 
x_4
end{bmatrix}=egin{bmatrix}
-cfrac{9}{34} & cfrac{9}{34} & -cfrac{3}{34} & cfrac{2}{17} 
cfrac{10}{17} & -cfrac{3}{34} & -cfrac{7}{51} & -cfrac{19}{51} 
cfrac{7}{17} & -cfrac{7}{17} & cfrac{8}{17} & -cfrac{5}{17} 
cfrac{1}{34} & -cfrac{1}{34} & cfrac{1}{102} &-cfrac{5}{51}
end{bmatrix}cdot egin{bmatrix}
3 
4 
2 
0
end{bmatrix}=egin{bmatrix}
-cfrac{9}{34}cdot 3 + cfrac{9}{34}cdot 4 -cfrac{3}{34}cdot 2  
cfrac{10}{17}cdot 3 -cfrac{3}{34}cdot 4 -cfrac{7}{51}cdot 2 
cfrac{7}{17}cdot 3 -cfrac{7}{17}cdot 4 + cfrac{8}{17}cdot 2 
cfrac{1}{34}cdot 3 -cfrac{1}{34}cdot 4 + cfrac{1}{102}cdot 2
end{bmatrix}=egin{bmatrix}
cfrac{3}{34} 
1cfrac{7}{51} 
cfrac{9}{17} 
-cfrac{1}{102}
end{bmatrix}

Ogólny wzór rozwiązywania układów równań tą metodą jest następujący:

gdzie:

Warto wspomnieć o macierzy ortogonalnej, która spełnia następującą równość:

Ponieważ iloczyn macierzy odwrotnej A-1 oraz macierzy A jest równy macierzy jednostkowej I, więc jeżeli dana macierz jest macierzą ortogonalną, to spełniony jest następujący warunek: