Monday, April 16, 2018

linear algebra - Calculate the dimension of a space of operators



This is a homework question.



First, consider the ring of real polynomials in n variables, Pn=R[x1,,xn] (n2), and let Sn act on Pn by automorphism (of algebra) by permutating the variables: σ(xi)=xσ(i)



For each i=1,,n1, let si be the transposition (i i+1) (so si(xi)=xi+1 and si(xi+1)=xi).




Finally, for each i, consider the operator Δi:PnPn given by
Δi(f)=fsi(f)xixi+1




(a) Show that Δi is a linear operator on Pn that satisfies the rule
Δi(fg)=Δi(f)g+si(f)Δi(g)
(b) Let Dn be the subalgebra generated by IdP,Δ1,,Δn1. Calculate the dimension of D3.





This is what I have so far: It is easy to check that Δi is a well-defined operator and that the rule above is satisfied. Moreover, in the basis {xα11xαnn}, we have the following (let's write x(α)=xα11xαnn for α=(α1,,αn))
Δi(x(α))=xα11xαi1i1ximin
(the sum is 0 if \alpha_i=\alpha_{i+1})



The problem is finding the dimension of D_3. I checked that \Delta_i^2=0, so D_3 is generated (as a vector space) by Id_P,\Delta_1,\Delta_1\Delta_2,\Delta_1\Delta_2\Delta_1,\ldots,\Delta_2,\Delta_2\Delta_1,\Delta_2\Delta_1\Delta_2,\ldots, but I couldn`t find any other good relation between \Delta_1 and \Delta_2 so obtain an upper bound for the dimension of D_3 (which I'm guessing is finite).


Answer



My previous answer was incorrect, here is the correct answer obtained by trying to correct the previous one :)



Claim: \Delta_1 \Delta_2 \Delta_1 = \Delta_2 \Delta_1 \Delta_2.




Note that the polynomial h = (x_1-x_2)(x_2-x_3)(x_3-x_1) has the property that s_i(h) = -h. It follows that \Delta_i h^2 p = h^2 \Delta_i p for any p.



A slightly mundane computation shows that
\Delta_1 \Delta_2 h(x_1,x_2,x_3) x_1^a x_2^b x_3^c = - x_1^a x_3^b x_2^{c+1}+x_1^a x_2^b x_3^{c+1}+x_1^{a+1} x_3^b x_2^c-\\x_1^{a+1} x_2^b x_3^c+x_3^a x_1^b x_2^{c+1}-x_2^a x_1^b x_3^{c+1}+x_2^a x_1^{b+1} x_3^c-x_3^a x_1^{b+1} x_2^c =: g(x_1,x_2,x_3).



Now, grouping terms with the same power of x_1 together and applying \Delta_2 shows (with some work) that:
\Delta_2 g(x_1,x_2,x_3) = x_1^a x_3^b x_2^c-x_1^a x_2^b x_3^c-x_3^a x_1^b x_2^c+x_2^a x_1^b x_3^c+x_3^a x_2^b x_1^c-x_2^a x_3^b x_1^c
A fully analogous argument show gives the same answer if we start with \Delta_2 \Delta_1 h(x_1,x_2,x_3) x_1^a x_2^b x_3^c. It follows that:
\Delta_1 \Delta_2 \Delta_1 h(x_1,x_2,x_3) x_1^a x_2^b x_3^c = \Delta_2 \Delta_1 \Delta_2 h(x_1,x_2,x_3) x_1^a x_2^b x_3^c.
It follows that \Delta_1 \Delta_2 \Delta_1 h p = \Delta_2 \Delta_1 \Delta_2 h p for any polynomial p. Thus, also \Delta_1 \Delta_2 \Delta_1 h^2 p = \Delta_2 \Delta_1 \Delta_2 h^2 p. We can now bring out h^2, and conclude that \Delta_1 \Delta_2 \Delta_1 and \Delta_2 \Delta_1 \Delta_2 .




It follows immediately (together with OP's work) that \Delta_1\Delta_2\Delta_1\Delta_2 = \Delta_2\Delta_1\Delta_2\Delta_1 =0. It remains to check that \mathrm{Id},\Delta_1,\Delta_2,\Delta_1 \Delta_2, \Delta_2 \Delta_1, \Delta_1\Delta_2\Delta_1 are linearly independent. They are, and it is enough to look at some low degree polynomials to check that. So, the sought dimension is 6.


No comments:

Post a Comment

analysis - Injection, making bijection

I have injection f \colon A \rightarrow B and I want to get bijection. Can I just resting codomain to f(A)? I know that every function i...