Zulip Chat Archive

Stream: new members

Topic: Prove that a MvPolynomial is symmetric


Xitai Jiang (Feb 20 2025 at 13:40):

I had been stuck to this point for a long time. I have that f (x, y, z) = f (y, x, z) = f (x, z, y) for all real numbers x, y, z, and how can I prove f is symmetric? Much thanks to all replies!

import Mathlib
theorem algebra_181 (P : MvPolynomial (Fin 3) ) (h :  x y z : ,
  (MvPolynomial.eval ![x, y, z] P = MvPolynomial.eval ![y, x, z] P  MvPolynomial.eval ![x, y, z] P = MvPolynomial.eval ![x, z, y] P)) : P.IsSymmetric := by
    simp [MvPolynomial.IsSymmetric]
    intro e
    fin_cases e
    simp
    all_goals
    sorry

Riccardo Brasca (Feb 20 2025 at 13:48):

I think doing it by brute force is going to be a pain. First of all one should prove that it is enough to prove symmetry wrt to a generating set of SnS_n (this can very well be already in the library), and then you are basically assuming that.

Philippe Duchon (Feb 20 2025 at 13:49):

For polynomials with real coefficients, one informal proof would go through the fact that coefficients can be recovered by taking derivatives and evaluating at 0; I don't know how difficult it would be to follow this path in Lean.


Last updated: May 02 2025 at 03:31 UTC