Error bounds for linear complementarity problems of Nekrasov matrices (Q483289): Difference between revisions

The linear complementarity problem associated with a given real square matrix \(M\in \mathbb{R}^{n\times n}\) and a vector \(q\in \mathbb{R}^{n}\) is the feasibility problem \[ LPC\left( M,q\right) \text{and }\,\, Mx+q\geq 0,\,x\geq 0,\,x^{T}\left( Mx+q\right) =0. \] \(M\) is called a \(P\)-matrix whenever its principal minors are positive or, equivalently, when \(LPC\left( M,q\right) \) has a unique solution for all \(q\in \mathbb{R}^{n}\). \(M\) is a nonsingular \(M\)-matrix if its inverse is nonnegative and all its off-diagonal entries are nonpositive. The comparison matrix of a complex square matrix \(M\in \mathbb{C} ^{n\times n}\) of entries \(m_{ij}\) is the real square matrix \(\mathcal{M} \left( M\right) \in \mathbb{R}^{n\times n}\) with diagonal entries \( \left| m_{ii}\right|\), \(i=1,\dots,n\), and off-diagonal entries \( -\left| m_{ij}\right|\), for all \(i\neq j\). \(M\in \mathbb{C} ^{n\times n}\) is said to be an \(H\)-matrix if \(\mathcal{M}\left( M\right) \) is a nonsingular \(M\)-matrix, and it is said to be strictly diagonally dominant by rows (SDD in short) if \(\left| m_{ii}\right| >\sum_{j\neq i}\left| m_{ij}\right| \) for all \(i\). There exists a wide literature on error bounds for \(LPC\left( M,q\right) \) when \(M\) is a real \(H\)-matrix or it belongs to some classes of \(P\)-matrices. This paper provides (in Theorems 3 and 5) error bounds for \(LPC\left( M,q\right) \) when \(M\) is either a Nekrasov matrix or a \(\Sigma \)-Nekrasov matrix. Both types of matrices are defined recursively. Any \(H\)-matrix is a Nekrasov matrix, and any Nekrasov matrix is an \(H\)-matrix and SDD. The new error bounds for \(LPC\left( M,q\right) \) are illustrated and compared with numerical examples.