Zero-family bounds on the full-slice product #
Pointwise qSDDOp bounds and averaged SDDOpRel bounds for the ordered fullSliceProductLeft
and reversed fullSliceProductRight factors against the zero family, each at most 1: the
counterpart of
MIPRE/Background/LIDT/MIPStarRE/LDT/Commutativity/Transport/FullSlice/ZeroBounds.lean in the
port of planning/c6b-plan.md (milestone M7, section "Port conventions").
The vendored proofs repeat, once per factor order, a Kronecker computation of
∑ (B_b A_a ⊗ I)^* (B_b A_a ⊗ I) as the placed total of the sandwich A_a B_b A_a. Here
it is proved once in the local algebra and placed by S.L, as
sum_ev_adjoint_self_leftTensor_mul_le_one below, which the evaluated-slice bounds of
Co/Commutativity/Main/Auxiliary/HEvalTransport.lean use as well. The vendored hypothesis
hnorm : strategy.state.IsNormalized is dropped: it is a theorem of the model
(VecState.ev_one_of_isNormalized).
New here #
sum_ev_adjoint_self_leftTensor_mul_le_one: for submeasurements A, B of 𝔓 with B
projective, ∑_{a,b} ev((B_b A_a ⊗ I)^* (B_b A_a ⊗ I)) ≤ 1.
Not ported #
Every declaration of the vendored file has a counterpart here.
References #
In LionSR/MIPStarRE at commit 507e8122, not in this repository:
references/ldt-paper/commutativity-G.texblueprint/src/chapter/ch08_commutativity.tex
For submeasurements A, B of the local algebra with B projective, the placed products
B_b A_a ⊗ I have total squared norm on the state at most 1: the sum
∑_{a,b} (B_b A_a)^* (B_b A_a) is the total ∑_a A_a (∑_b B_b) A_a of the sandwich
sandwichByOuterSubMeas A B, which is at most 1.
Questionwise, the ordered full-slice product has squared distance at most 1
from the zero family.
Questionwise, the reversed full-slice product has squared distance at most 1
from the zero family.
Averaging the ordered full-slice product against zero costs at most 1.
Averaging zero against the reversed full-slice product costs at most 1.