14.08.2013 Views

TH`ESE A NEW INVARIANT OF QUADRATIC LIE ALGEBRAS AND ...

TH`ESE A NEW INVARIANT OF QUADRATIC LIE ALGEBRAS AND ...

TH`ESE A NEW INVARIANT OF QUADRATIC LIE ALGEBRAS AND ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

2.4. 2-step nilpotent quadratic Lie algebras<br />

(1) Let A : g → g ′ be a Lie algebra isomorphism. Since h ∗ = Z(g) = Z(g ′ ) is stable by A then<br />

there exist linear maps A1 : h → h, A ′ 1 : h → h∗ and A2 : h ∗ → h ∗ such that<br />

A(x + f ) = A1(x) + A ′ 1(x) + A2( f ), ∀ x ∈ h, f ∈ h ∗ .<br />

It is obvious that A2 is an isomorphism of h ∗ . We show that A1 is also an isomorphism<br />

of h. Indeed, if there exists x0 ∈ h such that A1(x0) = 0 then 0 = [A(x0),g ′ ] ′ =<br />

A([x0,A −1 (g ′ )]) = A([x0,g]). It means that [x0,g] = 0. Since h ∗ = Z(g) then x0 = 0.<br />

Therefore A1 is an isomorphism of h.<br />

For all x,y ∈ h, f ,g ∈ h ∗ , one has<br />

A([x + f ,y + g]) = A(θ1(x,y)) = A2(θ1(x,y)).<br />

and [A(x + f ),A(y + g)] ′ = θ2(A1(x),A1(y)).<br />

Therefore, A2(θ1(x,y)) = θ2(A1(x),A1(y)) for all x,y ∈ h.<br />

Conversely, if there exist an isomorphism A1 of h and an isomorphism A2 of h ∗ such that<br />

A2(θ1(x,y)) = θ2(A1(x),A1(y)) for all x,y ∈ h, we define A : g → g ′ by A(x+ f ) = A1(x)+<br />

A2( f ) for all x + f ∈ g. Then it is easy to check that A is a Lie algebra isomorphism.<br />

(2) Asumme A : g → g ′ is an i-isomorphism then there exist A1 and A2 defined as in (1). Let<br />

x ∈ h, f ∈ h ∗ , one has<br />

B ′ (A(x),A( f )) = B(x, f ) ⇒ A2( f )(A1(x)) = f (x).<br />

Therefore, A2( f ) = f ◦ A −1<br />

1 , for all f ∈ h∗ .<br />

On the other hand, since A2(θ1(x,y)) = θ2(A1(x),A1(y)) we obtain<br />

θ1(x,y) = θ2(A1(x),A1(y)) ◦ A1, ∀ x,y ∈ h.<br />

Conversely, define A(x + f ) = A1(x) + f ◦ A −1<br />

1 , for all x ∈ h, f ∈ h∗ then A is an iisomorphism.<br />

Example 2.4.17. We keep the notations as above. Let g ′ be the T ∗ -extension of h by θ ′ = λθ<br />

where λ = 0. Then g and g ′ are i-isomorphic by A : g → g ′ defined by A(x + f ) = 1<br />

3√ λ x + 3√ λ f ,<br />

for all x + f ∈ g.<br />

For a non-degenerate cyclic skew-symmetric bilinear map θ of h, define the 3-form<br />

I(x,y,z) = θ(x,y)z, ∀ x,y,z ∈ h.<br />

Then I ∈ A 3 (h). The non-degenerate condition of θ is equivalent to ιx(I) = 0 for all x ∈ h\{0}.<br />

Conversely, let h be a vector space and I ∈ A 3 (h) such that ιx(I) = 0 for every non-zero vector<br />

x ∈ h. Define θ : h × h → h∗ by θ(x,y) = I(x,y,.), for all x,y ∈ h then θ is skew-symmetric<br />

and non-degenerate. Moreover, since I is alternating one has θ is cyclic and then we obtain a<br />

reduced 2SNQ-Lie algebra T ∗<br />

θ (h) defined by θ. Therefore, there is a one-to-one map from the<br />

set of all T ∗-extension of h onto the subset {I ∈ A 3 (h) | ιx(I) = 0, ∀ x ∈ h \ {0}}. We has a<br />

corollary of the above theorem as follows:<br />

58

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!