General solutions of linear matrix canonical differential equations with variable coefficients (Q844397): Difference between revisions

The linear matrix differential equation \[ \bigg(\frac{d}{dt}-A_1(t)\bigg)\bigg(\frac{d}{dt}-A_{2}(t)\bigg) \bigg(\dots\bigg(\bigg(\frac{d}{dt}- A_m(t)\bigg)\bigg)\dots\bigg)\bigg)X= F(t),\quad t\in [t_1,t_2]\subset(-\infty,+\infty),\tag{1} \] where \(A_i\) \((i=1,\dots,m),\) \(X,\) \(F\) are, respectively, \((n\times n)\)-, \((n\times l)\)-, \((n\times l)\)-matrices for \(n,l\in \mathbb N\), is investigated. The author introduces the notion of a regular matrix; namely, \(R\) is called a regular matrix if there exists an invertible matrix function \(\Phi\) such that \(\Phi(\int R(t)\,dt)=R(t)\Phi(\int R(t)\,dt)\). Some criteria for regularity are proved. For \(A_i,\) \(i=1,\dots,m,\) are regular, then a general solution of (1) is constructed. The results are applied to some special classes of differential equations, in particular, to equations of second order (including Gauss, Heun, Lamé, Legendre and Whittaker equations).