4.9 L’inverse d’un produit de matrices

Proposition 4.4 Soient \(\mathbf{A}, \mathbf{B}\in\mathbb{K}^{n\times n}\), où \(\mathbb{K}\) est un corps et \(n\) est un entier positif. Soit \(\mathbf{C} = \mathbf{A}\mathbf{B}\). Alors, \(\mathbf{C}\) est inversible si et seulement si \(\mathbf{A}\mathbf{B}\) est inversible.

Démonstration. Nous montrons la suffisance. La nécéssité est laissée en exercice.

Supposons que \(\mathbf{A}\) et \(\mathbf{B}\) soient toutes deux inversibles. Posons \(\mathbf{D} = \mathbf{B}^{-1}\mathbf{A}^{-1}\). Alors \[\begin{align*} \mathbf{D} \mathbf{C} & = (\mathbf{B}^{-1}\mathbf{A}^{-1})(\mathbf{A}\mathbf{B}) \\ & = \mathbf{B}^{-1}(\mathbf{A}^{-1}(\mathbf{A}\mathbf{B})) \\ & = \mathbf{B}^{-1}((\mathbf{A}^{-1}\mathbf{A})\mathbf{B}) \\ & = \mathbf{B}^{-1} (\mathbf{I}_n\mathbf{B}) = \mathbf{B}^{-1}\mathbf{B} \\ & = \mathbf{I}_n. \end{align*}\] De façon similaire, on peut vérifier que \[\mathbf{C} \mathbf{D} = \mathbf{I}_n.\] Ainsi, \(\mathbf{C}\) est inversible.

Plus généralement, si \(\mathbf{A}^{(1)},\ldots,\mathbf{A}^{(k)}\in \mathbb{K}^{n\times n}\) sont des matrices inversibles, alors \((\mathbf{A}^{(1)}\cdots \mathbf{A}^{(k)})^{-1}= {\mathbf{A}^{(k)}}^{-1}\cdots {\mathbf{A}^{(1)}}^{-1}.\)

Exemple 4.10 Soit \(\mathbf{A}= \begin{bmatrix} 1 & 0 \\ 0 & -1\end{bmatrix} \begin{bmatrix} 1 & 0 \\ -1 & 1\end{bmatrix} \begin{bmatrix} 1 & 3 \\ 0 & 1\end{bmatrix}.\) Alors, \(\mathbf{A} = \begin{bmatrix} 1 & 3\\1 & 2\end{bmatrix}\). De plus, \[\begin{align*} \mathbf{A}^{-1} & = \begin{bmatrix} 1 & 3 \\ 0 & 1\end{bmatrix}^{-1} \begin{bmatrix} 1 & 0 \\ -1 & 1\end{bmatrix}^{-1} \begin{bmatrix} 1 & 0 \\ 0 & -1\end{bmatrix}^{-1} \\ & = \begin{bmatrix} 1 & -3 \\ 0 & 1\end{bmatrix} \begin{bmatrix} 1 & 0 \\ 1 & 1\end{bmatrix} \begin{bmatrix} 1 & 0 \\ 0 & -1\end{bmatrix}. \end{align*}\]

Ainsi, \(\mathbf{A}^{-1} = \begin{bmatrix} -2 & 3\\1 & -1\end{bmatrix}\).

Les matrices obtenues ci-dessus sont aisées à obtenir puisqu’elles sont les inverses de matrices élémentaires.

4.9.1 Comment générer des matrices inversibles

Supposons que l’on nous demande de trouver quelques matrices inversibles \(3\times 3\) dont les inverses ne sons pas évidentes à trouver. Que ferions-nous?

Nous pourrions consulter des livres ou Internet, ou encore écrire quelques matrices \(3\times 3\) et utiliser la rédaction afin de trouver l’inverse de ces matrices, comme à la section précédente. Cependant, ces méthodes ne sont guère efficaces.

Le résultat clef nous permettent de générer une telle matrice inversible arbitraire est le suivant.

Proposition 4.5 Une matrice \(\mathbf{A} \in \mathbb{K}^{n\times n}\), où \(\mathbb{K}\) est un corps et \(n\) est un entier positif, est inversible si et seulement si \(\mathbf{A}\) est le produit de matrices élémentaires de \(\mathbb{K}^{n\times n}\).

Par exemple, \(\mathbf{A} = \begin{bmatrix} 1 & 2 \\ 3 & -1 \end{bmatrix}\) est inversible et peut s’écrire sous la forme \(\begin{bmatrix} 1&0\\3& 1\end{bmatrix} \begin{bmatrix} 1&0\\0& -7\end{bmatrix} \begin{bmatrix} 1&2\\0& 1\end{bmatrix}.\) La matrice \(\begin{bmatrix} 1&0\\3& 1\end{bmatrix}\) est la matrice élémentaire correspondant à l’opération élémentaire sur les lignes \(L_2 \leftarrow L_2 + 3L_1\). La matrice \(\begin{bmatrix} 1&0\\0& -7\end{bmatrix}\) est la matrice élémentaire correspondant à l’opération élémentaire sur les lignes \(L_2 \leftarrow -7L_2\). La matrice \(\begin{bmatrix} 1&2\\0& 1\end{bmatrix}\) est la matrice élémentaire correspondant à l’opération élémentaire sur les lignes \(L_1 \leftarrow L_1 + 2L_2\).

Avec le résultat précédant, il est possible de générer une matrice inversible arbitraire en commençant simplement avec une matrice élémentaire et en appliquant une suite arbitraire d’opérations élémentaires sur les lignes puisque la multiplication (à gauche) par des matrices élémentaires à le même effet que la d’opérations élémentaires sur les lignes correspondantes.

Démonstration. (Proposition 4.5) Rappelons que les matrices élémentaires sont inversibles et que le produit de matrices inversible est aussi inversible. Ainsi, si \(\mathbf{A}\) est le produit de matrices élémentaires, \(\mathbf{A}\) est inversible.

Nous montrons maitenant que si \(\mathbf{A}\) est inversible, alors on peut l’écrire en tant que produit de matrices élémentaires. Rappelons que si \(\mathbf{A}\) est inversible, alors la FER de la matrice \(\begin{bmatrix}\mathbf{A} & \mathbf{I}\end{bmatrix}\) est \(\begin{bmatrix}\mathbf{I} & \mathbf{A}^{-1}\end{bmatrix}\). Ainsi, il existe des matrices élémentaires \(\mathbf{M}^{(1)},\ldots,\mathbf{M}^{(k)}\) telles que \[\mathbf{M}^{(k)}\cdots \mathbf{M}^{(1)} \begin{bmatrix}\mathbf{A} & \mathbf{I}\end{bmatrix} = \begin{bmatrix}\mathbf{I} & \mathbf{A}^{-1}\end{bmatrix},\] ou de façon équivalente, \[\begin{bmatrix} \mathbf{M}^{(k)}\cdots \mathbf{M}^{(1)}\mathbf{A} & \mathbf{M}^{(k)}\cdots \mathbf{M}^{(1)}\end{bmatrix} = \begin{bmatrix}\mathbf{I} & \mathbf{A}^{-1}\end{bmatrix}.\] La comparaison des colonnes de la moitié droite de chaque côtés nous permet d’obtenir \[\mathbf{A}^{-1} = \mathbf{M}^{(k)}\cdots \mathbf{M}^{(1)}.\] Ainsi, \[\mathbf{A} = (\mathbf{A}^{-1})^{-1} = (\mathbf{M}^{(k)}\cdots \mathbf{M}^{(1)})^{-1} = {\mathbf{M}^{(1)}}^{-1}\cdots {\mathbf{M}^{(k)}}^{-1},\] qui est un produit de matrices élémentaires.

Exemple 4.11 Décomposons \(\mathbf{A} = \begin{bmatrix} 2 & 2 \\ 1 & 2 \end{bmatrix}\) en tant que produit de matrices élémentaires.

La discussion précédente suggère de réduire \(\begin{bmatrix}\mathbf{A} & \mathbf{I}\end{bmatrix}\) en en utilisant les opérations élémentaires sur les lignes. Mais nous n’avons pas à travailler avec la matrice \(\begin{bmatrix}\mathbf{A} & \mathbf{I}\end{bmatrix}\). Il suffit de connaître la suite des opérations élémentaires transformant \(\mathbf{A}\) en \(\mathbf{I}\).

Notons que \[\begin{align*} \begin{bmatrix} 2 & 1 \\ 2 & 2 \end{bmatrix} & \xrightarrow{L_1 \leftrightarrow L_2} \begin{bmatrix} 1 & 2 \\ 2 & 2 \end{bmatrix} \\ & \xrightarrow{L_2 \leftarrow L_2 - 2L_1} \begin{bmatrix} 1 & 2 \\ 0 & -2 \end{bmatrix} \\ & \xrightarrow{L_1 \leftarrow L_1 + L_2} \begin{bmatrix} 1 & 0 \\ 0 & -2 \end{bmatrix} \\ & \xrightarrow{L_2 \leftarrow -\frac{1}{2}L_2} \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix}. \end{align*}\] Ainsi, \[\begin{align*} \mathbf{A} & = \begin{bmatrix} 0 & 1 \\ 1 & 0\end{bmatrix}^{-1} \begin{bmatrix} 1 & 0 \\ -2 & 1 \end{bmatrix}^{-1} \begin{bmatrix} 1 & 1 \\ 0 & 1 \end{bmatrix}^{-1} \begin{bmatrix} 1 & 0 \\ 0 & -\frac{1}{2} \end{bmatrix}^{-1} \\ & = \begin{bmatrix} 0 & 1 \\ 1 & 0\end{bmatrix} \begin{bmatrix} 1 & 0 \\ 2 & 1 \end{bmatrix} \begin{bmatrix} 1 & -1 \\ 0 & 1 \end{bmatrix} \begin{bmatrix} 1 & 0 \\ 0 & -2 \end{bmatrix} \end{align*}\]

Exercices

  1. Déterminez l’inverse de chacun des produits suivants.

    1. \(\begin{bmatrix} 1 & -2 \\ 0 & 1 \end{bmatrix}\begin{bmatrix}0&1\\1&0 \end{bmatrix}\)

    2. \(\begin{bmatrix} 0 & 1 \\ -1 & 0 \end{bmatrix} \begin{bmatrix} 1 & 2 \\ 2 & 1 \end{bmatrix} \begin{bmatrix}3&0\\0&1 \end{bmatrix}\)

  2. Décomposez chacune des matrices suivantes en tant que produit de matrices élémentaires.

    1. \(\begin{bmatrix} 2 & 3 \\ 4 & 5 \end{bmatrix}\)

    2. \(\begin{bmatrix} 0 & 0 & 1 \\ 1 & 0 & 0 \\ 0 & 1 & 0 \end{bmatrix}\)

  3. Soient \(\mathbf{A}, \mathbf{B}\in \mathbb{K}^{n\times n}\), où \(\mathbb{K}\) est un corps. Démontrez que si \(\mathbf{A}\mathbf{B}\) est inversible, alors \(\mathbf{A}\) et \(\mathbf{B}\) sont inversibles.

Solutions

    1. \(\begin{bmatrix}0&1\\1&0 \end{bmatrix} \begin{bmatrix} 1 & 2 \\ 0 & 1 \end{bmatrix} = \begin{bmatrix} 0 & 1 \\ 1 & 2 \end{bmatrix}\)

    2. \(\begin{bmatrix}\frac{1}{3} &0\\0&1 \end{bmatrix} \begin{bmatrix} -\frac{1}{3} & \frac{2}{3} \\ \frac{2}{3} & -\frac{1}{3} \end{bmatrix} \begin{bmatrix} 0 & -1\\ 1 & 0 \end{bmatrix} = \begin{bmatrix} \frac{2}{9} & \frac{1}{9} \\ -\frac{1}{3} & -\frac{2}{3} \end{bmatrix}\)

    1. \(\begin{bmatrix} 1 & 0 \\ 2 & 1 \end{bmatrix} \begin{bmatrix} 1 & 3 \\ 0 & 1 \end{bmatrix} \begin{bmatrix} 2 & 0 \\ 0 & 1 \end{bmatrix}\)

    2. \(\begin{bmatrix} 1 & 0 & 0 \\ 0 & 0 & 1 \\ 0 & 1 & 0 \end{bmatrix} \begin{bmatrix} 0 & 0 & 1 \\ 0 & 1 & 0 \\ 1 & 0 & 0 \end{bmatrix}\)

  1. Posons \(\mathbf{C}=\mathbf{A}\mathbf{B}.\) Puisque \(\mathbf{C}\) est inversible, il existe \(\mathbf{D} \in \mathbb{K}^{n\times n}\) telle que \(\mathbf{D}\mathbf{C} = \mathbf{C}\mathbf{D} = \mathbf{I}_n.\) Ainsi, \[\begin{align*} \mathbf{D}\mathbf{A}\mathbf{B} & = \mathbf{I}_n \\ \mathbf{A}\mathbf{B}\mathbf{D} & = \mathbf{I}_n, \end{align*}\] ce qui implique \[\begin{align*} (\mathbf{D}\mathbf{A})\mathbf{B} & = \mathbf{I}_n \\ \mathbf{A}(\mathbf{B}\mathbf{D}) & = \mathbf{I}_n. \end{align*}\] Ainsi, \(\mathbf{D}\mathbf{A}\) est l’inverse à gauche de \(\mathbf{B}\) et \(\mathbf{B}\mathbf{D}\) est l’inverse à droite de \(\mathbf{A}.\) D’après la proposition 4.2, \(\mathbf{A}\) et \(\mathbf{B}\) ont donc un inverse à droite et un inverse à gauche, respectivement. Ainsi, \(\mathbf{A}\) et \(\mathbf{B}\) sont toutes les deux inversibles.