claimroutineinferred_dep
Source-release cost f/u via incoming cap
Releasing a source costs by the incoming cap
source-release cost in the comparison
Claim·§7.3 Mean comparisons and the cost of releasing a source·inside lemMean comparison scale factors and release cost
Statement (paper)
Let be a left support relation with incoming degree at most , and let be a common right discrepancy. Then where the source channel in the last norm is . Hence, when a source is released, the second root-discrepancy sum, after eliminating by the left incoming cap and using , costs , while the first sum costs only a constant factor via the right row cap. A projected term costs .
In plain English
If a left support relation has incoming degree at most and is a common right discrepancy, then summing over with weight and over with is bounded by , with source channel . So releasing a source costs only in norm power.
Why it's here
Each is counted by at most values of , so the -sum collapses to a factor and the weights change from to relative to the norm. This gives the release cost in the parent lemma, applied to the second root-discrepancy sum, while the first sum needs only a constant factor from the right row cap. The bound adds to the scale-factor bounds of the sibling claims and needs no independence.
Classified routine:
A direct double-counting bound: since each z has at most K_c incoming x, the x-sum is bounded using X_u\le 1, and the remaining step is unfolding the definition of the weighted norm with weight f. This is bookkeeping inside the parent mean-comparison lemma.
Uses
Used by
Sources
PaperOpen
- Section
- §7.3 Mean comparisons and the cost of releasing a source
- Page
- —
- Label
—- Line
sections/estimation.tex:258
Leannot aligned
Phase 2 aligns this node to a Lean declaration.
Friction
- infoinferred_depoutline: "Source-release cost f/u via incoming cap" rests on claim-lem-est-mean-comparison-1