Root-row pencils and the polynomial root-column argument #
A finite homogeneous-linear system becomes a shared Pencil by multiplying its
output row by the central polynomial variable. If that matrix is nilpotent,
1 - T has a polynomial right inverse. Eliminating the internal entries of its
root column gives q = 1 + X * C(x) * q. The coefficients of this polynomial
are x^n, and their eventual vanishing proves nilpotence of x.
A polynomial right resolvent forces nilpotence, also in a noncommutative ring and without assuming that the ring is nontrivial.
Reindexing square matrices preserves products, zero, and one.
Equations
Instances For
The root is none; only its row contains X.
Equations
- S.matrix a none none = Polynomial.X * KoetheCounterexample.Linearization.edge a S.head
- S.matrix a none (some j_2) = Polynomial.X * KoetheCounterexample.Linearization.edge a (S.out j_2)
- S.matrix a (some i_2) none = KoetheCounterexample.Linearization.edge a (S.input i_2)
- S.matrix a (some i_2) (some j_2) = KoetheCounterexample.Linearization.edge a (S.step i_2 j_2)
Instances For
Enumeration of the states with the shared root index 0.
Equations
- S.indexEquiv = (finSuccEquiv (Fintype.card S.State)).trans (Fintype.equivFin S.State).symm.optionCongr
Instances For
Exactly the shared pencil API, with no scalar/identity edges.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Nilpotence transfers from the shared, finitely indexed pencil to the same matrix indexed by the root and internal states.
Eliminate the internal entries of the root column of a polynomial right inverse. Nilpotence is only used to obtain this right inverse.
Each finite linearization gives a single-row pencil whose nilpotence implies nilpotence of the represented element.