File size: 41,002 Bytes
4a56cd6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 | import OAI.Algorithms.SubexponentialSAT.LeanCompat
import OAI.Algorithms.SubexponentialSAT.Truncation
import OAI.Algorithms.SubexponentialSAT.Moments
import OAI.Algorithms.SubexponentialSAT.Separation
import OAI.Algorithms.SubexponentialSAT.GlobalPolynomial
import OAI.Algorithms.SubexponentialSAT.JetCRT
import OAI.Algorithms.SubexponentialSAT.LiftRing
import OAI.Algorithms.SubexponentialSAT.NilpotentDefect
import OAI.Algorithms.SubexponentialSAT.LayoutSpec
import OAI.Algorithms.SubexponentialSAT.Incidence
import OAI.Algorithms.SubexponentialSAT.TruncatedInverse
import OAI.Algorithms.SubexponentialSAT.ActualLiftRing
namespace OAI.Algorithms
/-!
# Exact counting by separated symbolic moments
This module defines the separated finite sums, rather than assuming they
return the target count. Canonical deletion is applied before coefficient
extraction and before evaluating U at one. Theorems make the algebraic
hypotheses explicit; they do not claim an executable bit-complexity bound.
-/
namespace SubexponentialSAT.Counting
noncomputable section
open scoped BigOperators
abbrev Poly (ι R : Type*) [CommSemiring R] := Truncation.Poly ι R
section Local
variable {ι Ω O R : Type*} {X Y : O → Type*}
[CommSemiring R] [Fintype ι] [Fintype Ω] [Fintype O]
[∀ o, Fintype (X o)] [∀ o, Fintype (Y o)]
/-- A group polynomial after deleting precisely the forbidden monomials. -/
def group (μ D : ℕ) (P : Ω → Poly ι R) (z : Ω) : Poly ι R :=
Truncation.canonical μ D (D + μ) (P z)
/-- The canonical product of the two separately truncated groups. -/
def groupProduct (μ D : ℕ) (P Q : Ω → Poly ι R) (z : Ω) : Poly ι R :=
Truncation.canonical μ D (D + μ) (group μ D P z * group μ D Q z)
/-- The actual two inner sums of equation (4.8), followed by canonical
reduction of the result. The input equivalence enumerates every assignment
exactly once by outer data and two fibers. -/
def separatedMoment (μ D : ℕ) (e : Ω ≃ Σ o, X o × Y o)
(r₀ : ∀ o, X o) (r₁ : ∀ o, Y o) (P Q : Ω → Poly ι R)
(c : O → R) (w₀ : ∀ o, X o → R) (w₁ : ∀ o, Y o → R) : Poly ι R :=
Truncation.canonical μ D (D + μ)
(∑ o, WeightedSupport.scalarHom (c o) *
(∑ x : X o, WeightedSupport.scalarHom (w₀ o x) *
group μ D Q (e.symm ⟨o, x, r₁ o⟩)) *
(∑ y : Y o, WeightedSupport.scalarHom (w₁ o y) *
group μ D P (e.symm ⟨o, r₀ o, y⟩)))
/-- Extract the X monomial while retaining both polynomial variables U,Z. -/
def extractX (a : ι →₀ ℕ) : Poly ι R →+ Polynomial (Polynomial R) :=
Truncation.mapCoeffs (Truncation.mapCoeffs (MvPolynomial.coeffAddMonoidHom a))
/-- Equation (4.7), with all moment sums actually given by the separated
construction. U=1 is applied only after canonical reduction and extraction. -/
def separatedJet (μ D : ℕ) (e : Ω ≃ Σ o, X o × Y o)
(r₀ : ∀ o, X o) (r₁ : ∀ o, Y o) (P Q : Ω → Poly ι R)
(c : (ι →₀ ℕ) → O → R)
(w₀ : (ι →₀ ℕ) → ∀ o, X o → R)
(w₁ : (ι →₀ ℕ) → ∀ o, Y o → R) : Polynomial R :=
∑ a ∈ Moments.retained (D + μ),
(extractX a (separatedMoment μ D e r₀ r₁ P Q (c a) (w₀ a) (w₁ a))).eval 1
omit [Fintype ι] [Fintype Ω] in
/-- The product used in moment extraction equals the canonical reduction of
the full product, even though the two groups were truncated independently. -/
theorem groupProduct_eq (μ D : ℕ) (P Q : Ω → Poly ι R) (z : Ω) :
groupProduct μ D P Q z = Truncation.canonical μ D (D + μ) (P z * Q z) := by
exact Truncation.canonical_mul μ D (D + μ) (P z) (Q z)
omit [Fintype ι] [Fintype Ω] in
/-- Exact group invariance passes through canonical truncation. -/
theorem group_congr (μ D : ℕ) (P : Ω → Poly ι R) {z z' : Ω}
(h : P z = P z') : group μ D P z = group μ D P z' := congrArg (Truncation.canonical μ D (D + μ)) h
omit [Fintype ι] in
/-- No target-count equality is assumed: the unsimplified separated sum is
proved equal to the full symbolic moment by finite distributivity. -/
theorem separatedMoment_eq_canonical_sum (μ D : ℕ)
(e : Ω ≃ Σ o, X o × Y o) (r₀ : ∀ o, X o) (r₁ : ∀ o, Y o)
(P Q : Ω → Poly ι R) (w : Ω → R)
(c : O → R) (w₀ : ∀ o, X o → R) (w₁ : ∀ o, Y o → R)
(hP : ∀ o x x' y, P (e.symm ⟨o, x, y⟩) = P (e.symm ⟨o, x', y⟩))
(hQ : ∀ o x y y', Q (e.symm ⟨o, x, y⟩) = Q (e.symm ⟨o, x, y'⟩))
(hw : ∀ o x y, w (e.symm ⟨o, x, y⟩) = c o * w₀ o x * w₁ o y) :
separatedMoment μ D e r₀ r₁ P Q c w₀ w₁ =
Truncation.canonical μ D (D + μ)
(∑ z, WeightedSupport.scalarHom (w z) * group μ D P z * group μ D Q z) := by
unfold separatedMoment
congr 1
symm
apply Separation.weighted_separation_equiv e r₀ r₁
(group μ D P) (group μ D Q) (fun z => WeightedSupport.scalarHom (w z))
(fun o => WeightedSupport.scalarHom (c o))
(fun o x => WeightedSupport.scalarHom (w₀ o x))
(fun o y => WeightedSupport.scalarHom (w₁ o y))
· intro o x x' y
exact group_congr μ D P (hP o x x' y)
· intro o x y y'
exact group_congr μ D Q (hQ o x y y')
· intro o x y
rw [hw, map_mul, map_mul]
omit [Fintype ι] in
/-- Moving canonical reduction through the finite sums and scalar weights
identifies the separated output with the full moment of canonical products. -/
theorem separatedMoment_eq_sum (μ D : ℕ)
(e : Ω ≃ Σ o, X o × Y o) (r₀ : ∀ o, X o) (r₁ : ∀ o, Y o)
(P Q : Ω → Poly ι R) (w : Ω → R)
(c : O → R) (w₀ : ∀ o, X o → R) (w₁ : ∀ o, Y o → R)
(hP : ∀ o x x' y, P (e.symm ⟨o, x, y⟩) = P (e.symm ⟨o, x', y⟩))
(hQ : ∀ o x y y', Q (e.symm ⟨o, x, y⟩) = Q (e.symm ⟨o, x, y'⟩))
(hw : ∀ o x y, w (e.symm ⟨o, x, y⟩) = c o * w₀ o x * w₁ o y) :
separatedMoment μ D e r₀ r₁ P Q c w₀ w₁ =
∑ z, WeightedSupport.scalarHom (w z) * groupProduct μ D P Q z := by
rw [separatedMoment_eq_canonical_sum μ D e r₀ r₁ P Q w c w₀ w₁ hP hQ hw,
Truncation.canonical_sum]
apply Finset.sum_congr rfl
intro z hz
rw [mul_assoc, Truncation.canonical_scalar_mul]
rfl
/-- Every retained coefficient really has support in the finite moment set. -/
theorem canonical_support (μ D N j r : ℕ) (P : Poly ι R) :
(((Truncation.canonical μ D N P).coeff j).coeff r).support ⊆
Moments.retained N := by
intro a ha
rw [Moments.mem_retained]
by_contra h
have hc := MvPolynomial.mem_support_iff.mp ha
apply hc
rw [Truncation.coeff_canonical]
split_ifs
· simp [Truncation.coeff_xTrunc, h]
· simp
omit [Fintype ι] [Fintype Ω] in
/-- Exact substitution of the canonical group product is the desired local
truncation; the two hypotheses are support filtration and nilpotence. -/
theorem groupProduct_substitution (μ D : ℕ) (P Q : Ω → Poly ι R)
(v : Ω → ι → R)
(hP : ∀ z, WeightedSupport.Admissible (P z))
(hQ : ∀ z, WeightedSupport.Admissible (Q z))
(hhigh : ∀ z j, μ ≤ j → (Truncation.substX (v z) (P z * Q z)).coeff j = 0)
(z : Ω) :
Truncation.evalXU (v z) (groupProduct μ D P Q z) =
Truncation.trunc D (Truncation.evalXU (v z) (P z * Q z)) := by
exact Truncation.truncated_groups_substitution (v z) μ D (P z) (Q z)
(hP z) (hQ z) (hhigh z)
/-- The defined separated-jet construction is the sum of the correct local
truncations. Its hypotheses concern input invariance, scalar weights, support,
and nilpotence; no equality with the desired jet is assumed. -/
theorem separatedJet_eq_sum (μ D : ℕ)
(e : Ω ≃ Σ o, X o × Y o) (r₀ : ∀ o, X o) (r₁ : ∀ o, Y o)
(P Q : Ω → Poly ι R) (v : Ω → ι → R)
(c : (ι →₀ ℕ) → O → R)
(w₀ : (ι →₀ ℕ) → ∀ o, X o → R)
(w₁ : (ι →₀ ℕ) → ∀ o, Y o → R)
(hP : ∀ o x x' y, P (e.symm ⟨o, x, y⟩) = P (e.symm ⟨o, x', y⟩))
(hQ : ∀ o x y y', Q (e.symm ⟨o, x, y⟩) = Q (e.symm ⟨o, x, y'⟩))
(hw : ∀ a o x y, Moments.weight a (v (e.symm ⟨o, x, y⟩)) =
c a o * w₀ a o x * w₁ a o y)
(hPa : ∀ z, WeightedSupport.Admissible (P z))
(hQa : ∀ z, WeightedSupport.Admissible (Q z))
(hhigh : ∀ z j, μ ≤ j → (Truncation.substX (v z) (P z * Q z)).coeff j = 0) :
separatedJet μ D e r₀ r₁ P Q c w₀ w₁ =
∑ z, Truncation.trunc D (Truncation.evalXU (v z) (P z * Q z)) := by
classical
calc
separatedJet μ D e r₀ r₁ P Q c w₀ w₁ =
∑ a ∈ Moments.retained (D + μ),
(extractX a (∑ z, WeightedSupport.scalarHom (Moments.weight a (v z)) *
groupProduct μ D P Q z)).eval 1 := by
unfold separatedJet
apply Finset.sum_congr rfl
intro a ha
rw [separatedMoment_eq_sum μ D e r₀ r₁ P Q
(fun z => Moments.weight a (v z)) (c a) (w₀ a) (w₁ a) hP hQ (hw a)]
_ = ∑ z, Truncation.evalXU (v z) (groupProduct μ D P Q z) := by
exact Moments.sum_retained_nested_moments_eval_one_eq_sum_evalXU
(D + μ) (groupProduct μ D P Q) v
(fun z j r => canonical_support μ D (D + μ) j r _)
_ = _ := by
apply Finset.sum_congr rfl
intro z hz
exact groupProduct_substitution μ D P Q v hPa hQa hhigh z
end Local
section Rows
variable {V E Ω R : Type*} [CommRing R] [Fintype V] {σ : E → Type*}
/-- The coefficient of Z^j in a shifted row, kept as an actual linear form
in the formal input variables. -/
def rowCoefficient (row : V → Polynomial R) (θ : R) (j : ℕ) : MvPolynomial V R :=
WeightedSupport.linearForm Finset.univ
(fun v => ((row v).comp (Polynomial.X + Polynomial.C θ)).coeff j)
/-- The manuscript's incidence argument, with the positive-Z expansion
constructed directly from the polynomial row. -/
def rowArgument (row : V → Polynomial R) (θ β : R) (r : ℕ) : Poly V R :=
WeightedSupport.incidenceArgument β (rowCoefficient row θ 0)
((Finset.range (r + 1)).erase 0) (rowCoefficient row θ)
theorem rowArgument_admissible (row : V → Polynomial R) (θ β : R) (r : ℕ) :
WeightedSupport.Admissible (rowArgument row θ β r) := by
apply WeightedSupport.admissible_incidence_linearForms
intro j hj
exact Nat.pos_of_ne_zero (Finset.mem_erase.mp hj).1
/-- Substitution cancels beta and reconstructs the shifted scalar row. -/
theorem evalXU_rowArgument (row : V → Polynomial R) (x : V → R)
(θ β : R) (r : ℕ) (hrow : ∀ v, (row v).natDegree ≤ r) :
Truncation.evalXU x (rowArgument row θ β r) =
(GlobalPolynomial.rowValue row x).comp (Polynomial.X + Polynomial.C θ) := by
rw [rowArgument, Truncation.evalXU_incidenceArgument]
exact (GlobalPolynomial.shifted_rowValue_expansion row x θ r hrow).symm
/-- A full symbolic factor group. Its dependence on the actual input z is
confined to beta; every rowCoefficient remains a formal polynomial. -/
def rowGroup (s : Finset E) (table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) (z : Ω) : Poly V R :=
∏ e ∈ s, MvPolynomial.eval₂ (WeightedSupport.scalarHom (ι := V))
(fun i => rowArgument (row e i) θ (β z e i) r) (table e)
theorem rowGroup_admissible (s : Finset E) (table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) (z : Ω) :
WeightedSupport.Admissible (rowGroup s table row θ r β z) := by
apply WeightedSupport.admissible_factor_product
intro e he i
exact rowArgument_admissible (row e i) θ (β z e i) r
theorem rowGroup_congr (s : Finset E) (table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) {z z' : Ω}
(hβ : ∀ e ∈ s, ∀ i, β z e i = β z' e i) :
rowGroup s table row θ r β z = rowGroup s table row θ r β z' := by
apply Finset.prod_congr rfl
intro e he
congr 1
funext i
rw [hβ e he i]
/-- Evaluation of an entire symbolic group is its shifted factor product. -/
theorem evalXU_rowGroup (s : Finset E) (table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) (z : Ω) (x : V → R)
(hrow : ∀ e ∈ s, ∀ i v, (row e i v).natDegree ≤ r) :
Truncation.evalXU x (rowGroup s table row θ r β z) =
∏ e ∈ s, GlobalPolynomial.compose (table e)
(fun i => (GlobalPolynomial.rowValue (row e i) x).comp
(Polynomial.X + Polynomial.C θ)) := by
rw [rowGroup, Truncation.evalXU_factor_product]
apply Finset.prod_congr rfl
intro e he
congr 1
funext i
exact evalXU_rowArgument (row e i) x θ (β z e i) r (hrow e he i)
/-- Disjoint factor groups multiply to the full group. -/
theorem rowGroup_union [DecidableEq E] (s t : Finset E) (hst : Disjoint s t)
(table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) (z : Ω) :
rowGroup (s ∪ t) table row θ r β z =
rowGroup s table row θ r β z * rowGroup t table row θ r β z :=
Finset.prod_union hst
/-- Prime-power divisibility kills every discarded U coefficient for the
actual row/table construction; it is not assumed for the final product. -/
theorem rowGroup_highU (s : Finset E) (table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) (z : Ω) (x : V → R)
(p : R) (μ : ℕ) (hnil : p ^ μ = 0)
(hβ : ∀ e ∈ s, ∀ i, p ∣ MvPolynomial.eval x (rowCoefficient (row e i) θ 0) - β z e i)
(j : ℕ) (hj : μ ≤ j) :
(Truncation.substX x (rowGroup s table row θ r β z)).coeff j = 0 := by
exact NilpotentDefect.highU_factor_product x p μ hnil s table (β z)
(fun e i => rowCoefficient (row e i) θ 0)
(fun _ _ => (Finset.range (r + 1)).erase 0)
(fun e i => rowCoefficient (row e i) θ) hβ j hj
end Rows
section GlobalRows
variable {V E Ω R : Type*} [CommRing R] [Fintype V] [Fintype E] [DecidableEq E]
[Fintype Ω] {σ : E → Type*}
/-- The two actual symbolic factor groups give exactly the formal expansion
of H at the chosen node after X substitution and U=1. -/
theorem sum_evalXU_rowGroups (s t : Finset E) (hst : Disjoint s t)
(hcover : s ∪ t = Finset.univ)
(table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) (v : Ω → V → R)
(hrow : ∀ e i w, (row e i w).natDegree ≤ r) :
(∑ z, Truncation.evalXU (v z)
(rowGroup s table row θ r β z * rowGroup t table row θ r β z)) =
(GlobalPolynomial.globalPolynomial table
(fun z e i => GlobalPolynomial.rowValue (row e i) (v z))).comp
(Polynomial.X + Polynomial.C θ) := by
classical
rw [GlobalPolynomial.comp_globalPolynomial]
unfold GlobalPolynomial.globalPolynomial
apply Finset.sum_congr rfl
intro z hz
rw [← rowGroup_union s t hst, hcover]
exact evalXU_rowGroup Finset.univ table row θ r β z (v z)
(fun e _ i w => hrow e i w)
/-- The correct local jet expressed as the sum of the truncated group
substitutions, before the moment-separation optimization is invoked. -/
theorem sum_trunc_evalXU_rowGroups (_μ D : ℕ)
(s t : Finset E) (hst : Disjoint s t) (hcover : s ∪ t = Finset.univ)
(table : ∀ e, MvPolynomial (σ e) R)
(row : ∀ e, σ e → V → Polynomial R) (θ : R) (r : ℕ)
(β : Ω → ∀ e, σ e → R) (v : Ω → V → R)
(hrow : ∀ e i w, (row e i w).natDegree ≤ r) :
(∑ z, Truncation.trunc D (Truncation.evalXU (v z)
(rowGroup s table row θ r β z * rowGroup t table row θ r β z))) =
Truncation.trunc D
((GlobalPolynomial.globalPolynomial table
(fun z e i => GlobalPolynomial.rowValue (row e i) (v z))).comp
(Polynomial.X + Polynomial.C θ)) := by
have h := congrArg (Truncation.truncAddHom (A := R) D)
(sum_evalXU_rowGroups s t hst hcover table row θ r β v hrow)
rw [map_sum] at h
exact h
end GlobalRows
section TruncationBridge
variable {R : Type*} [CommRing R]
/-- Canonical coefficient deletion agrees with monic reduction modulo Z^D. -/
theorem trunc_eq_modByMonic (D : ℕ) (f : Polynomial R) :
Truncation.trunc D f = f %ₘ (Polynomial.X ^ D) := by
nontriviality R
have hdiv : (Polynomial.X : Polynomial R) ^ D ∣ f - Truncation.trunc D f := by
rw [Polynomial.X_pow_dvd_iff]
intro n hn
simp [Truncation.coeff_trunc, hn]
have hself : Truncation.trunc D f %ₘ (Polynomial.X ^ D) = Truncation.trunc D f := by
apply (Polynomial.modByMonic_eq_self_iff (Polynomial.monic_X.pow D)).2
rw [Polynomial.degree_X_pow]
apply (Polynomial.degree_lt_iff_coeff_zero _ D).2
intro n hn
simp [Truncation.coeff_trunc, Nat.not_lt.mpr hn]
exact ((Polynomial.modByMonic_eq_of_dvd_sub (Polynomial.monic_X.pow D) hdiv).trans hself).symm
theorem jet_eq_trunc (θ : R) (D : ℕ) (f : Polynomial R) :
JetCRT.jet θ D f = Truncation.trunc D (f.comp (Polynomial.X + Polynomial.C θ)) := by
exact (trunc_eq_modByMonic D _).symm
end TruncationBridge
section LayoutCounting
variable {p k Δ : ℕ} [Fact p.Prime]
{V F K R : Type*} [Fintype V] [DecidableEq V] [Fintype F] [DecidableEq F]
[Field K] [Fintype K] [Algebra (ZMod p) K] [FiniteDimensional (ZMod p) K]
[CommRing R] {P : Factor.Pattern V F} {c : LayoutSpec.Constants p k Δ}
{L : LayoutSpec.Layout (E := K) P c}
/-- The actual residue lift used as beta at each incidence. -/
def layoutBeta (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue) (t : Fin c.Q)
(x : V → ZMod p) (e : F) (i : {v // v ∈ P.scope e}) : R :=
(A.liftedRow lift hLift t e i).beta s x
/-- Factor colors determine the two concrete groups. -/
def colorGroup (L : LayoutSpec.Layout (E := K) P c) (t : Fin c.Q) (g : Bool) : Finset F :=
Finset.univ.filter (fun e => L.color e t = g)
omit [DecidableEq F] [Fintype K] [FiniteDimensional (ZMod p) K] in
theorem colorGroup_disjoint (L : LayoutSpec.Layout (E := K) P c) (t : Fin c.Q) :
Disjoint (colorGroup L t false) (colorGroup L t true) := by
apply Finset.disjoint_left.mpr
intro e he hf
simp only [colorGroup, Finset.mem_filter, Finset.mem_univ, true_and] at he hf
rw [he] at hf
cases hf
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
theorem colorGroup_union (L : LayoutSpec.Layout (E := K) P c) (t : Fin c.Q) :
colorGroup L t false ∪ colorGroup L t true = Finset.univ := by
ext e
simp only [colorGroup, Finset.mem_union, Finset.mem_filter, Finset.mem_univ, true_and]
cases L.color e t <;> simp
/-- The actual symbolic groups determined by a lifted layout. -/
def layoutGroup (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R)
(t : Fin c.Q) (g : Bool) : (V → ZMod p) → Poly V R :=
rowGroup (colorGroup L t g) table A.row (A.theta t) c.R
(layoutBeta A lift hLift s t)
omit [DecidableEq F] [Fintype K] [FiniteDimensional (ZMod p) K] in
theorem layoutGroup_left_invariant (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R) (t : Fin c.Q)
(o : L.Outer t) (x x' : L.Left t o) (y : L.Right t o) :
layoutGroup A lift hLift s table t false (L.reconstruct t o x y) =
layoutGroup A lift hLift s table t false (L.reconstruct t o x' y) := by
apply rowGroup_congr
intro e he i
exact A.beta_reconstruct_left lift hLift s t e i
(Finset.mem_filter.mp he).2 o x x' y
omit [DecidableEq F] [Fintype K] [FiniteDimensional (ZMod p) K] in
theorem layoutGroup_right_invariant (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R) (t : Fin c.Q)
(o : L.Outer t) (x : L.Left t o) (y y' : L.Right t o) :
layoutGroup A lift hLift s table t true (L.reconstruct t o x y) =
layoutGroup A lift hLift s table t true (L.reconstruct t o x y') := by
apply rowGroup_congr
intro e he i
exact A.beta_reconstruct_right lift hLift s t e i
(Finset.mem_filter.mp he).2 o x y y'
omit [DecidableEq V] in
/-- The constant coefficient used in the symbolic argument is exactly the
actual lifted row value, before selecting beta. -/
theorem eval_rowCoefficient_zero (row : V → Polynomial R) (x : V → R) (θ : R) :
MvPolynomial.eval x (rowCoefficient row θ 0) = ∑ w, (row w).eval θ * x w := by
classical
simp [rowCoefficient, WeightedSupport.linearForm, MvPolynomial.eval_monomial,
Polynomial.coeff_zero_eq_eval_zero, Polynomial.eval_comp, mul_comm]
omit [DecidableEq F] [Fintype K] [FiniteDimensional (ZMod p) K] in
/-- The local p-divisibility hypothesis is derived from the actual residue
kernel and the chosen section. -/
theorem layoutBeta_defect (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R))
(t : Fin c.Q) (x : V → ZMod p) (e : F) (i : {v // v ∈ P.scope e}) :
(p : R) ∣ MvPolynomial.eval (fun w => lift (x w))
(rowCoefficient (A.row e i) (A.theta t) 0) - layoutBeta A lift hLift s t x e i := by
rw [eval_rowCoefficient_zero]
exact (A.liftedRow lift hLift t e i).value_sub_beta_dvd s (p : R) hker x
omit [DecidableEq F] [Fintype K] [FiniteDimensional (ZMod p) K] in
/-- A concrete point in a nonempty coset, chosen independently of all weights. -/
def cosetRepresentative {𝔽 M : Type*} [Field 𝔽] [AddCommGroup M] [Module 𝔽 M]
(U : Submodule 𝔽 M) (q : M ⧸ U) : CosetCardinality.CosetFiber U q :=
⟨Classical.choose (Submodule.Quotient.mk_surjective (p := U) q),
Classical.choose_spec (Submodule.Quotient.mk_surjective (p := U) q)⟩
def layoutOuterWeight (L : LayoutSpec.Layout (E := K) P c) (lift : ZMod p → R)
(t : Fin c.Q) (a : V →₀ ℕ) (o : L.Outer t) : R :=
∏ i : ↥((L.support t false ∪ L.support t true)ᶜ), lift (o.1 i) ^ a i.val
def layoutLeftWeight (L : LayoutSpec.Layout (E := K) P c) (lift : ZMod p → R)
(t : Fin c.Q) (a : V →₀ ℕ) (o : L.Outer t) (x : L.Left t o) : R :=
∏ i : L.support t false, lift (x.val i) ^ a i.val
def layoutRightWeight (L : LayoutSpec.Layout (E := K) P c) (lift : ZMod p → R)
(t : Fin c.Q) (a : V →₀ ℕ) (o : L.Outer t) (y : L.Right t o) : R :=
∏ i : L.support t true, lift (y.val i) ^ a i.val
local instance leftFintype (t : Fin c.Q) (o : L.Outer t) : Fintype (L.Left t o) := by
classical
unfold LayoutSpec.Layout.Left CosetCardinality.CosetFiber
infer_instance
local instance rightFintype (t : Fin c.Q) (o : L.Outer t) : Fintype (L.Right t o) := by
classical
unfold LayoutSpec.Layout.Right CosetCardinality.CosetFiber
infer_instance
/-- The complete separated local computation, with groups, representatives,
weights, and outer/inner domains supplied by the actual layout. -/
def layoutJet (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R)
(μ D : ℕ) (t : Fin c.Q) : Polynomial R := by
classical
exact separatedJet μ D (L.coordinateOuterFiberEquiv t)
(fun o => cosetRepresentative (L.localK t false) o.2.1)
(fun o => cosetRepresentative (L.localK t true) o.2.2)
(layoutGroup A lift hLift s table t false)
(layoutGroup A lift hLift s table t true)
(layoutOuterWeight L lift t) (layoutLeftWeight L lift t) (layoutRightWeight L lift t)
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
/-- All semantic hypotheses of the separated-jet theorem are discharged by
the actual layout, residue kernel, and nilpotence. -/
theorem layoutJet_eq (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R)
(μ D : ℕ) (hnil : (p : R) ^ μ = 0)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R))
(t : Fin c.Q) :
layoutJet A lift hLift s table μ D t =
JetCRT.jet (A.theta t) D
(GlobalPolynomial.rowGlobalPolynomial table A.row lift) := by
classical
calc
layoutJet A lift hLift s table μ D t =
∑ x : V → ZMod p, Truncation.trunc D
(Truncation.evalXU (fun w => lift (x w))
(layoutGroup A lift hLift s table t false x *
layoutGroup A lift hLift s table t true x)) := by
unfold layoutJet
refine separatedJet_eq_sum (ι := V) (Ω := V → ZMod p) (O := L.Outer t)
(X := L.Left t) (Y := L.Right t) (R := R) μ D
(L.coordinateOuterFiberEquiv t)
(fun o => cosetRepresentative (L.localK t false) o.2.1)
(fun o => cosetRepresentative (L.localK t true) o.2.2)
(layoutGroup A lift hLift s table t false)
(layoutGroup A lift hLift s table t true) (fun x w => lift (x w))
(layoutOuterWeight L lift t) (layoutLeftWeight L lift t)
(layoutRightWeight L lift t) ?_ ?_ ?_ ?_ ?_ ?_
· exact layoutGroup_left_invariant A lift hLift s table t
· exact layoutGroup_right_invariant A lift hLift s table t
· intro a o x y
exact Separation.momentWeight_coordinateReconstruct
(L.support t false) (L.support t true) (L.support_disjoint t)
(L.localK t false) (L.localK t true) o x y a lift
· intro x
exact rowGroup_admissible _ _ _ _ _ _ _
· intro x
exact rowGroup_admissible _ _ _ _ _ _ _
· intro x j hj
change (Truncation.substX (fun w => lift (x w))
(rowGroup (colorGroup L t false) table A.row (A.theta t) c.R
(layoutBeta A lift hLift s t) x *
rowGroup (colorGroup L t true) table A.row (A.theta t) c.R
(layoutBeta A lift hLift s t) x)).coeff j = 0
rw [← rowGroup_union _ _ (colorGroup_disjoint L t), colorGroup_union]
exact rowGroup_highU Finset.univ table A.row (A.theta t) c.R
(layoutBeta A lift hLift s t) x (fun w => lift (x w)) (p : R) μ hnil
(fun e _ i => layoutBeta_defect A lift hLift s hker t x e i) j hj
_ = JetCRT.jet (A.theta t) D
(GlobalPolynomial.rowGlobalPolynomial table A.row lift) := by
rw [jet_eq_trunc]
exact sum_trunc_evalXU_rowGroups μ D _ _ (colorGroup_disjoint L t)
(colorGroup_union L t) table A.row (A.theta t) c.R
(layoutBeta A lift hLift s t) (fun x w => lift (x w)) A.row_degree
omit [DecidableEq F] [Fintype K] [FiniteDimensional (ZMod p) K] in
/-- Unit differences of lifted nodes follow from residue separation and the
principal nilpotent kernel; they are not an extra interpolation hypothesis. -/
theorem theta_unit_differences (A : LayoutSpec.LiftedLayout L R)
(s : CosetInvariance.ResidueSection A.residue) (μ : ℕ)
(hnil : (p : R) ^ μ = 0)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R)) :
Pairwise (fun i j => IsUnit (A.theta i - A.theta j)) := by
intro i j hij
apply NilpotentDefect.isUnit_of_residue_ne_zero A.residue
(fun z => ⟨s.lift z, s.residue_lift z⟩) (p : R) μ hnil
· intro z hz
rw [← Ideal.mem_span_singleton, ← hker]
exact hz
· rw [map_sub, A.residue_theta, A.residue_theta]
intro h
exact hij (L.node_injective (sub_eq_zero.mp h))
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
/-- The concrete finite-sum CRT reconstruction fed by the actual separated
jet computation. The supplied polynomials are algebraic Bezout certificates. -/
def reconstructLayout (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R)
(μ D : ℕ) (u : Fin c.Q → Polynomial R) : Polynomial R :=
JetCRT.reconstructJetsWithBezout A.theta D u (layoutJet A lift hLift s table μ D)
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
/-- CRT recovers H from the defined separated jets. The degree hypothesis is
a polynomial degree condition, discharged below for interpolated factors. -/
theorem reconstructLayout_eq (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R)
(μ D : ℕ) (hnil : (p : R) ^ μ = 0)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R))
(u v : Fin c.Q → Polynomial R)
(hbez : ∀ t, u t * JetCRT.cofactor A.theta D t +
v t * JetCRT.nodePower (A.theta t) D = 1)
(hdegree : (GlobalPolynomial.rowGlobalPolynomial table A.row lift).natDegree < c.Q * D) :
reconstructLayout A lift hLift s table μ D u =
GlobalPolynomial.rowGlobalPolynomial table A.row lift := by
nontriviality R
apply JetCRT.reconstructJetsWithBezout_eq A.theta D
(theta_unit_differences A s μ hnil hker) u v _ hbez
· apply Polynomial.degree_le_natDegree.trans_lt
exact WithBot.coe_lt_coe.mpr (by simpa using hdegree)
· intro t
rw [layoutJet_eq A lift hLift s table μ D hnil hker t]
exact (TruncatedInverse.jet_modByMonic _ _ _).symm
/-- Each inverse of a node difference is computed by lifting its residue-field
inverse and applying a finite nilpotent geometric correction. -/
def differenceInverse (A : LayoutSpec.LiftedLayout L R)
(s : CosetInvariance.ResidueSection A.residue) (μ : ℕ)
(i j : Fin c.Q) : R :=
NilpotentDefect.residueInverse A.residue s.lift μ (A.theta i - A.theta j)
omit [DecidableEq F] [Fintype K] [FiniteDimensional (ZMod p) K] in
theorem differenceInverse_mul (A : LayoutSpec.LiftedLayout L R)
(s : CosetInvariance.ResidueSection A.residue) (μ : ℕ)
(hnil : (p : R) ^ μ = 0)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R))
(i j : Fin c.Q) (hij : i ≠ j) :
differenceInverse A s μ i j * (A.theta i - A.theta j) = 1 := by
rw [mul_comm]
apply NilpotentDefect.mul_residueInverse A.residue s.lift s.residue_lift (p : R) μ hnil
· intro z hz
rw [← Ideal.mem_span_singleton, ← hker]
exact hz
· rw [map_sub, A.residue_theta, A.residue_theta]
intro h
exact hij (L.node_injective (sub_eq_zero.mp h))
/-- Explicit finite geometric CRT reconstruction of the separated jets.
There are no supplied Bezout polynomials or inverse certificates. -/
def explicitReconstruction (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R)
(μ D : ℕ) : Polynomial R :=
TruncatedInverse.reconstruct A.theta D
(TruncatedInverse.constantInverse D (differenceInverse A s μ))
(layoutJet A lift hLift s table μ D)
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
theorem explicitReconstruction_eq (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(s : CosetInvariance.ResidueSection A.residue)
(table : ∀ e, MvPolynomial {v // v ∈ P.scope e} R)
(μ D : ℕ) (hD : 0 < D) (hnil : (p : R) ^ μ = 0)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R))
(hdegree : (GlobalPolynomial.rowGlobalPolynomial table A.row lift).natDegree < c.Q * D) :
explicitReconstruction A lift hLift s table μ D =
GlobalPolynomial.rowGlobalPolynomial table A.row lift := by
nontriviality R
let _ : Nonempty (Fin c.Q) := ⟨⟨0, c.Q_pos⟩⟩
apply TruncatedInverse.reconstruct_eq A.theta D hD
(theta_unit_differences A s μ hnil hker)
· exact TruncatedInverse.constantInverse_correct A.theta D _
(differenceInverse_mul A s μ hnil hker)
· simpa using hdegree
· intro t
rw [layoutJet_eq A lift hLift s table μ D hnil hker t, jet_eq_trunc]
exact TruncatedInverse.trunc_idempotent D _
end LayoutCounting
section InterpolatedCounting
variable {p k Δ : ℕ} [Fact p.Prime]
{V F K R : Type*} [Fintype V] [DecidableEq V] [Fintype F] [DecidableEq F]
[Field K] [Fintype K] [Algebra (ZMod p) K] [FiniteDimensional (ZMod p) K]
[CommRing R] (S : Factor.System V F (ZMod p)) {c : LayoutSpec.Constants p k Δ}
{L : LayoutSpec.Layout (E := K) S.toPattern c}
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
omit [DecidableEq F] in
/-- The strict interpolation degree budget follows from the actual bounded
factor pattern and the individual degrees of the explicit table interpolants. -/
theorem interpolated_degree_lt (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R) (hu : TableInterpolation.UnitDifferences lift)
(hD : 0 < Fintype.card V) :
(GlobalPolynomial.interpolatedGlobalPolynomial S lift hu A.row).natDegree <
c.Q * Fintype.card V := by
have hm : Fintype.card F ≤ Δ * Fintype.card V := by
apply Incidence.card_factors_le S.scope Δ L.bounded.nonempty_scope
intro w
exact L.bounded.incidence_le w
have hd := GlobalPolynomial.natDegree_interpolatedGlobalPolynomial_le S lift hu A.row
k c.R L.bounded.arity_le A.row_degree
calc
(GlobalPolynomial.interpolatedGlobalPolynomial S lift hu A.row).natDegree ≤
Fintype.card F * k * (p - 1) * c.R := by simpa only [ZMod.card] using hd
_ ≤ (Δ * Fintype.card V) * k * (p - 1) * c.R :=
Nat.mul_le_mul_right c.R (Nat.mul_le_mul_right (p - 1) (Nat.mul_le_mul_right k hm))
_ = Fintype.card V * (Δ * k * (p - 1) * c.R) := by ac_rfl
_ < Fintype.card V * c.Q := Nat.mul_lt_mul_of_pos_left c.degree_budget hD
_ = c.Q * Fintype.card V := Nat.mul_comm _ _
omit [DecidableEq F] in
/-- The actual count value produced by separated jets and explicit CRT. -/
def recoveredValue (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(hu : TableInterpolation.UnitDifferences lift)
(s : CosetInvariance.ResidueSection A.residue)
(μ : ℕ) (u : Fin c.Q → Polynomial R) : R :=
(reconstructLayout A lift hLift s (GlobalPolynomial.interpolatedTable S lift hu)
μ (Fintype.card V) u).eval 0
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
/-- The layout-to-counting correctness theorem. Every local jet is the
separated construction above; the target equality is a conclusion. -/
theorem recoveredValue_eq_count (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(hu : TableInterpolation.UnitDifferences lift)
(s : CosetInvariance.ResidueSection A.residue)
(μ : ℕ) (hnil : (p : R) ^ μ = 0)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R))
(hD : 0 < Fintype.card V) (u v : Fin c.Q → Polynomial R)
(hbez : ∀ t, u t * JetCRT.cofactor A.theta (Fintype.card V) t +
v t * JetCRT.nodePower (A.theta t) (Fintype.card V) = 1) :
recoveredValue S A lift hLift hu s μ u = (Factor.count S : R) := by
unfold recoveredValue
rw [reconstructLayout_eq A lift hLift s _ μ (Fintype.card V) hnil hker u v hbez
(interpolated_degree_lt S A lift hu hD)]
exact GlobalPolynomial.eval_zero_interpolatedGlobalPolynomial S lift hu A.row A.row_zero
/-- The same count computation using finite geometric CRT throughout. -/
def explicitRecoveredValue (A : LayoutSpec.LiftedLayout L R) (lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(hu : TableInterpolation.UnitDifferences lift)
(s : CosetInvariance.ResidueSection A.residue) (μ : ℕ) : R :=
(explicitReconstruction A lift hLift s (GlobalPolynomial.interpolatedTable S lift hu)
μ (Fintype.card V)).eval 0
omit [Fintype K] [FiniteDimensional (ZMod p) K] in
theorem explicitRecoveredValue_eq_count (A : LayoutSpec.LiftedLayout L R)
(lift : ZMod p → R)
(hLift : ∀ z, A.residue (lift z) = algebraMap (ZMod p) K z)
(hu : TableInterpolation.UnitDifferences lift)
(s : CosetInvariance.ResidueSection A.residue)
(μ : ℕ) (hnil : (p : R) ^ μ = 0)
(hker : RingHom.ker A.residue = Ideal.span ({(p : R)} : Set R))
(hD : 0 < Fintype.card V) :
explicitRecoveredValue S A lift hLift hu s μ = (Factor.count S : R) := by
unfold explicitRecoveredValue
rw [explicitReconstruction_eq A lift hLift s _ μ (Fintype.card V) hD hnil hker
(interpolated_degree_lt S A lift hu hD)]
exact GlobalPolynomial.eval_zero_interpolatedGlobalPolynomial S lift hu A.row A.row_zero
omit [DecidableEq F] in
/-- The chosen prime-power modulus strictly exceeds every possible count. -/
theorem factor_count_lt_prime_power : Factor.count S < p ^ (Fintype.card V + 1) := by
calc
Factor.count S ≤ p ^ Fintype.card V := by
simpa only [ZMod.card] using Factor.count_le_pow S
_ < p ^ (Fintype.card V + 1) :=
Nat.pow_lt_pow_right (Fact.out : Nat.Prime p).one_lt (Nat.lt_succ_self _)
end InterpolatedCounting
section ActualCounting
variable {p k Δ : ℕ} [Fact p.Prime]
{V F : Type*} [Fintype V] [DecidableEq V] [Fintype F] [DecidableEq F]
(S : Factor.System V F (ZMod p)) {c : LayoutSpec.Constants p k Δ}
(L : LayoutSpec.Layout (E := FiniteField.E p c.h) S.toPattern c)
/-- The completely instantiated ring-valued output: the actual monic-lift
quotient, integer representatives, lifted layout, residue section, separated
moments, and finite geometric CRT are all constructed. -/
def actualRingValue : ActualLiftRing.Ring p c.h (Fintype.card V + 1) :=
explicitRecoveredValue S
(ActualLiftRing.liftedLayout L (Fintype.card V + 1) (Nat.succ_pos _))
(ActualLiftRing.coordinateLift p c.h (Fintype.card V + 1))
(ActualLiftRing.rho_coordinateLift p c.h (Fintype.card V + 1) (Nat.succ_pos _))
(ActualLiftRing.coordinateLift_unitDifferences p c.h (Fintype.card V + 1))
(ActualLiftRing.residueSection p c.h (Fintype.card V + 1) (Nat.succ_pos _))
(Fintype.card V + 1)
/-- Every algebraic side condition is discharged for the actual prime-power
ring. Only the supplied layout and nonempty coordinate domain are required. -/
theorem actualRingValue_eq_count (hD : 0 < Fintype.card V) :
actualRingValue S L =
(Factor.count S : ActualLiftRing.Ring p c.h (Fintype.card V + 1)) := by
unfold actualRingValue
exact explicitRecoveredValue_eq_count S
(ActualLiftRing.liftedLayout L (Fintype.card V + 1) (Nat.succ_pos _))
(ActualLiftRing.coordinateLift p c.h (Fintype.card V + 1))
(ActualLiftRing.rho_coordinateLift p c.h (Fintype.card V + 1) (Nat.succ_pos _))
(ActualLiftRing.coordinateLift_unitDifferences p c.h (Fintype.card V + 1))
(ActualLiftRing.residueSection p c.h (Fintype.card V + 1) (Nat.succ_pos _))
(Fintype.card V + 1)
(ActualLiftRing.prime_power_zero p c.h (Fintype.card V + 1))
(ActualLiftRing.rho_ker p c.h (Fintype.card V + 1) (Nat.succ_pos _)) hD
/-- Decode the actual ring output as a natural number using its canonical
constant coefficient. This definition does not call the target factor count. -/
def actualCount : ℕ :=
ActualLiftRing.decodeCount p c.h (Fintype.card V + 1) (Nat.succ_pos _)
(actualRingValue S L)
/-- Exact natural-number counting from an interpolation layout. The count is
recovered, not merely tested modulo p or compared for zero. No ring, lift,
coefficient, inverse, or desired-result equality is a hypothesis. -/
theorem actualCount_eq_count (hD : 0 < Fintype.card V) :
actualCount S L = Factor.count S := by
rw [actualCount, actualRingValue_eq_count S L hD]
exact ActualLiftRing.decodeCount_natCast p c.h (Fintype.card V + 1) (Nat.succ_pos _)
(Nat.ne_of_gt c.h_pos) (Factor.count S) (factor_count_lt_prime_power S)
/-- The resulting decision test has exactly the intended satisfying-assignment
semantics. Complexity and a concrete machine realization are separate. -/
theorem actualCount_pos_iff (hD : 0 < Fintype.card V) :
0 < actualCount S L ↔ ∃ x : V → ZMod p, Factor.Satisfies S x := by
rw [actualCount_eq_count S L hD]
exact Factor.count_pos_iff S
end ActualCounting
end
end SubexponentialSAT.Counting
end OAI.Algorithms
|