Solving a kind of restricted matrix equations and Cramer rule (Q1763280): Difference between revisions

The authors establish determinantal formulas for solutions of the restricted matrix equations \(AX=D\) \((R(X)\subset R(A^{k_1}))\), \(XB=D\) \((N(X)\supset N(B^{k_2}))\), \(AXB=D\) \((R(X)\subset R(A^{k_1}),N(X)\supset N(B^{k_2}))\). The complex matrices \(A\), \(B\), \(D\) are respectively \(n\times n\), \(m\times m\), \(n\times m\). One has Ind\((A)=k_1\), Ind\((B)=k_2\) where Ind\((A)=\min \{ k\in {\mathbb N}\mid \text{rank} (A^k)= \text{rank} (A^{k+1})\}\). Here \(N(X)\) and \(R(X)\) denote the null and the range space of the matrix \(X\). The authors show that the unique solutions of the above equations equal respectively \(A_dD\), \(DB_d\), \(A_dDB_d\) where \(A_d\) is the Drazin inverse of \(A\), i.e. the unique matrix \(X\) satisfying the conditions \(A^{k_1+1}X=A^{k_1}\), \(XAX=X\), \(AX=XA\).