Graded decomposition (Lemma 6.3(1)) #
positive-deformations.tex: "Every finite graded A₀-module is a sum of
shifts of the M_i. […] Decompose its underlying ungraded module into normalized
modules M_i. Write the projections and inclusions as p_a and j_a, so
1_X = Σ_a p_a j_a. Taking homogeneous degree-zero parts gives
1_X = Σ_{a,r} (p_a)_r (j_a)_{-r}."
Deviation in the proof: instead of the locality of End_gr(X) for a graded
indecomposable X, we split off one shifted M_W at a time from a graded
idempotent e (exists_summand). Some composite (p_a)_r (j_b)_{-r} of equal
level is nonzero, hence invertible by eq:normalized-hom-degrees (exists_inv);
otherwise the level sums E_r satisfy E_r E_s = 0 for r ≤ s and force e = 0.
The result (exists_gdecomp) is a finite family of graded inclusions and
projections with Σ inc_a pr_a = 1, which is the input of extVanish_of_decomp.
A nonzero degree-zero map between evaluated modules is invertible.
A graded A₀-linear endomorphism of degree zero.
Instances For
One shifted M_W splits off a nonzero graded idempotent.
A graded decomposition of a graded idempotent e into shifted evaluated
modules: e = Σ_a inc_a pr_a with inc_a : M_{W a}⟨s_a⟩ → X and
pr_a : X → M_{W a}⟨s_a⟩ graded A₀-linear of degree zero.
- μ : Type
- fin : Fintype self.μ
- W : self.μ → V
- s : self.μ → ℤ
- inc (a : self.μ) : CatAlg.Rep D.F0 (meshFam Q c (self.W a)) →ₗ[k] X
- pr (a : self.μ) : X →ₗ[k] CatAlg.Rep D.F0 (meshFam Q c (self.W a))
Instances For
Lemma 6.3(1): every finite graded A₀-module is a finite sum of shifts of
evaluated modules.