2025-06-08

1154: Non-Degenerate Hermitian Matrix

<The previous article in this series | The table of contents of this series | The next article in this series>

definition of non-degenerate Hermitian matrix

Topics


About: matrices space

The table of contents of this article


Starting Context



Target Context


  • The reader will have a definition of non-degenerate Hermitian matrix.

Orientation


There is a list of definitions discussed so far in this site.

There is a list of propositions discussed so far in this site.


Main Body


1: Structured Description


Here is the rules of Structured Description.

Entities:
M: { the n×n Hermitian matrices }
//

Conditions:
v{ the n×1 complex matrices }
(
w{ the n×1 complex matrices }(wMv=0)v=0
)
//

The condition, v{ the n×1 complex matrices } such that v0(w{ the n×1 complex matrices }(wMv0)), is an equivalent condition (called "the 1st alternative condition"), as will be seen in Note.

The condition, f:CnCn,v(wvMw), where Cn is the covectors space of Cn, L(Cn:C), is a 'vectors spaces - linear morphisms' isomorphism, is an equivalent condition (called "the 2nd alternative condition"), as will be seen in Note.

The condition, M can be diagonalized with all nonzero diagonal components with a unitary matrix, is an equivalent condition (called "the 3nd alternative condition"), as will be seen in Note.


2: Note


Let us see that the 1st alternative condition is an equivalent condition.

Let us suppose the condition of this definition.

Let v0 be any.

If there was no w such that wMv0, for each w, wMv=0, which would imply v=0, a contradiction.

So, there is a w.

Let us suppose that 1st alternative condition.

Let v be any.

Let us suppose that for each w, wMv=0.

If v0, there would be a w such that wMv0, a contradiction against that wMv=0 for each w.

So, v=0.

Let us see that the 2nd alternative condition is an equivalent condition.

1st, let us see that f(v) is indeed in Cn.

For each wCn, f(v)(w)=vMwC. For each w,wCn and each r,rC, f(v)(rw+rw)=vM(rw+rw)=rvMw+rvMw=rf(v)(w)+rf(v)(w), which means that f(v) is a multilinear map.

So, f(v)Cn.

Let us suppose the condition of this definition.

Let us see that f is linear.

f(rv+rv)(w)=(rv+rv)Mw=rvMw+rvMw=rf(v)(w)+rf(v)(w)=(rf(v)+rf(v))(w), which implies that f(rv+rv)=rf(v)+rf(v).

Let us see that f is injective.

Let v,vCn be any such that vv.

Let us suppose that f(v)=f(v). For each wCn, f(v)(w)f(v)(w)=0, f(vv)(w)=(vv)Mw=0, so, ((vv)Mw)=0=0, but the left hand side is wM(vv)=wM(vv), so, wM(vv)=0, which would imply that vv=0, so, v=v, a contradiction.

So, f is injective.

So, f is a 'vectors spaces - linear morphisms' isomorphism, by the proposition that any linear injection between any same-finite-dimensional vectors spaces is a 'vectors spaces - linear morphisms' isomorphism: any covectors space is known to be same-dimensional with the original space.

Let us suppose the 2nd alternative condition.

Let vCn be any such that v0.

There is a wCn such that vMw0, because as f is injective, f(v)f(0), which means that there is a w such that f(v)(w)0 (otherwise, f(v)=f(0)), and f(v)(w)=vMw0, so, (vMw)0=0, but the left hand side is wMv=wMv, so, wMv0.

So, the 1st alternative condition is satisfied, which implies that the condition of this definition is satisfied.

Let us see that the 3rd alternative condition is an equivalent condition.

Note that it is known that any Hermitian matrix, M, can be diagonalized by a unitary matrix, U, as UMU, with all real diagonal components. The issues is whether there is no zero diagonal component.

Let us suppose the condition of this definition.

wMv=wU1UMUU1v=wU1(UMU)U1v, by the proposition that for any invertible complex matrix, the inverse of the Hermitian conjugate of the matrix is the Hermitian conjugate of the inverse of the matrix, =(U1w)(UMU)U1v, by the proposition that the Hermitian conjugate of the product of any complex matrices is the product of the Hermitian conjugates of the constituents in the reverse order.

If UMU had a zero component, a nonzero v:=U1v can be chosen such that UMUU1v=0 (let only the corresponding component of U1v be nonzero), and (U1w)(UMU)U1v=0 for each w:=U1w; but v=Uv would be nonzero because U was unitary and wMv=(U1w)(UMU)U1v would be 0 for each w, a contradiction.

So, UMU cannot have any zero component.

Let us suppose the 3rd alternative condition.

As before, wMv=(U1w)(UMU)U1v.

Let us suppose that for each wCn, wMv=0.

wMv=(U1w)(UMU)U1v=0, and U1w can be chosen such that only the j-th component of (U1w)(UMU) is nonzero (let only the corresponding component of U1w be nonzero), which implies that the j-th component of U1v is 0, so, U1v=0, so, v=0.

So, the condition of this definition is satisfied.


References


<The previous article in this series | The table of contents of this series | The next article in this series>