A wild quiver in linear systems theory (Q1091988): Difference between revisions

The main objective is to construct a canonical form for the action \(\alpha\) : ((T,L),(A,B))\(\to (T^{-1}AT,T^{-1}BL)\) of the group \(Gl_ n(K)\times Gl_ n(K)\) of linear coordinate transformations in the state (T) and input space (L) on the space \(U_{n,m}\) of control systems \((A,B)\in K^{n\times m}\times K^{n\times m}\) with n different eigenvalues (K algebraically closed field). The authors also classify the orbits of \(\alpha\) by discrete indices, find their dimensions and determine the stabilizer subgroup, with respect to \(\alpha\), of \((A,B)\in U_{n,m}\) up to isomorphy. Combinatorics are used with success to analyze the scaling action \(\beta\) : ((D,E),B)\(\to D^{-1}BE\) of the group of diagonal matrices \(D_ n(K)\times D_ m(K)\) on \(K^{n\times m}\) and, in particular, to construct a canonical form for this action. The study of the action \(\beta\) is important not just in its own, but it is also used to solve the main problem.