Reason: P8.0: closeness of the realized control to A and of the realized mean-field flow to S on a high-probability observation-measurable event, extracted from the cost bound (CB) and (I') through the block cascade; input to the close-record mass for the synthetic copy.
Statement
Data. Adopt the setting, notation and standing hypotheses of the asymptotic lower bound theorem for the recentred N-agent cost, and with them those of the ledger lemma and of the block cascade lemma on which it rests. In particular: the affine-controlled transition-rate family(β0,β1) on l states with compact convex control set A⊆Rm and control bound R, and its transition-rate familyβ with aggregate state drift b, state-Lipschitz constant Λb and constant K1; the horizon T>0; the stationary mean-field triple(S,A,P) with x0=S0 in the probability simplexΔl, so that the standing hypothesis S∗=S of the block cascade lemma holds; and the family, indexed by the natural numbersN≥1, of solutions of the controlled N-agent dynamics on [0,T], the N-th of them carried by an N-agent driving system(Ω,F,P) with regular event Ω0 and observation filtration (Gt)t∈[0,T], together with all the standing hypotheses imposed there, among them (I′) with the bound κ♯ and (CB) with the bound J♯. As there, the probability of the driving system is written P, the stationary co-state always carrying a time subscript, and the superscript N is suppressed throughout.
of claim 4 of the extended good-set stopping-time lemma. None of α^, Φ, Y, E, CS and none of the conclusions of that lemma's claim 4 involves the thresholds δ, θout, εY and cE fixed in its setting; wherever that claim is invoked below, those four thresholds may therefore be given the value 1.
Adopt finally the notion of an admissible parameter vector π=(T0,ε1,λc,λo,δcl,q0,ϵ,η,ϱ), the absorption condition of clause (a) of the lower bound theorem, and the real number Z♯(π) defined there. For an admissible π and a natural number N≥1, write K for the number of blocks, t0,…,tK for the grid, Lk for the levels, σ(k) for the anchored good-set clocks, G0,…,GK for the good sets, D0,…,DK−1 for the leave events, ΔkE=Emin(σ(k),tk+1)−Etk for the block energy increments, Λ⋆ for the constant, and
Z=Nk=0∑K−1E[1GkΔkE]
for the tracked energy, all of these being the objects the block cascade lemma forms from π for the N-th solution; and write P=∑k=0K−1P(Dk) and Υlev=∑k=0K−1Lk−2 as in the ledger lemma. Here E is the expectation, 1D is the function equal to 1 on a set D and 0 off it, ∣⋅∣ is the Euclidean norm, ∫[a,b]⋅ds is the Lebesgue integral over a compact interval, taken to be 0 when a=b, exp is the exponential function and x1/2=x is the nonnegative square root of a nonnegative real x.
Then the following hold.
1. (Energy accrued while tracked, and the escape probability.) Let π be admissible and let N≥1. Then
E[1GKET]≤NZ,P(Ω∖GK)=P≤NΛ⋆ΥlevZ,
all three quantities being finite.
2. (Pathwise bounds carried by the control energy.) For every N≥1, every ω∈Ω and every t∈[0,T],
Neither bound involves the parameter vector, and neither excludes an exceptional set.
3. (A parameter vector with a uniform tracked energy bound.) There are an admissible parameter vector π∙ satisfying the absorption condition of clause (a) of the lower bound theorem, a real number Z♯≥0 and a natural number N1, all three determined by the common data and by the bounds κ♯ and J♯ and not by the individual solutions, such that Z♯=Z♯(π∙) and such that for every N≥N1 the tracked energy of the N-th solution formed from π∙ satisfies Z≤Z♯.
4. (The good event.) Let π∙, Z♯ and N1 be as in claim 3, let Λ⋆, Υlev, K and, for each N≥1, the good set GK be those formed from π∙, and set
The constants Cctl, Cflw and Cesc depend only on the common data and on the bounds κ♯ and J♯; the tolerance ε♭ enters the three pathwise bounds only through the factor ε♭−1, enters the probability bound directly, and enters the threshold N♭ through the requirement ε♭N♭≥Cesc. The three bounds of claim 4 hold at every time of [0,T] simultaneously, and are of the orders N−1, N−1/2 and N−1/2. No hypothesis bounding E[∣st∣p] or E[∣at∣p] uniformly in N at positive times is used: the closeness of the realized control to A and of the realized mean-field flow to S is extracted from the cost bound (CB) together with the hypothesis (I′) on the initial fourth moment, through the block cascade, and only on the event ΩN♭, off which no bound on either quantity is asserted.
Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.