Skip to main content

Part 3· Chapter 12

Chapter 12 — Three Computational Axes

Prerequisites: Chapter 11 (Cech framework), Chapter 6 (Door).


12.1 Overview

Once the world's constitution is fixed, invariants are extracted via three computational axes:

AxisObjectGeometric meaning
Axis 1H(Kα;A)H^*(K_\alpha;A)Intrinsic topology of the fruit
Axis 2H(Kα,Lα;A)H^*(K_\alpha,L_\alpha;A)Relative topology of the fruit--door pair
Axis 3H(Lα;A)H^*(L_\alpha;A) + persistence barcodeTopology and fractal structure of the door set

The sequential protocol extends:

WtFtΣtAxis 1Axis 2Axis 3.\mathcal{W}_t\Rightarrow\mathfrak{F}_t\Rightarrow\Sigma_t\Rightarrow\text{Axis 1}\Rightarrow\text{Axis 2}\Rightarrow\text{Axis 3}.

Each axis depends only on the preceding ones.


12.2 Axis 1: Fruit Topology H(Kα;A)H^*(K_\alpha;A)

Definition 12.1 (Fruit complex)

Given fruit FFtF\in\mathfrak{F}_t, construct the clique complex KαK_\alpha of the induced graph Gt[F]\mathcal{G}_t[F]: a kk-simplex for each (k+1)(k+1)-clique.

Theorem 12.1 (Finiteness of the nerve)

dim(Nerve(Kα))F1\dim(\mathrm{Nerve}(K_\alpha))\le|F|-1.

Proof. The maximum clique size in FF is at most F|F|. \square

Computation

Hk(Kα;A)=kerDKk/imDKk1H^k(K_\alpha;A)=\ker D_K^k/\mathrm{im}\,D_K^{k-1}

where DKk=(Bk+1K)TD_K^k=(B_{k+1}^K)^T is the coboundary matrix built from the boundary matrices of KαK_\alpha.

Lemma 12.2 (U(1)U(1) case)

For G=U(1)G=U(1) and abelian coefficients A=ZA=\mathbb{Z}: Hˇ(Kα;U(1))Hom(H(Kα;Z),U(1))\check H^*(K_\alpha;U(1))\cong\mathrm{Hom}(H_*(K_\alpha;\mathbb{Z}),U(1)) by the universal coefficient theorem.


12.3 Axis 2: Pair Cohomology H(Kα,Lα;A)H^*(K_\alpha,L_\alpha;A)

Definition 12.2 (Simplicial pair)

  • KαK_\alpha: the fruit complex (Definition 12.1).
  • LαKαL_\alpha\subset K_\alpha: the door subcomplex, generated by all simplices containing at least one door node.

Theorem 12.3 (Long exact sequence of a pair)

For each degree kk, there is a long exact sequence:

Hk(Kα,Lα;A)JkHk(Kα;A)RkHk(Lα;A)δHk+1(Kα,Lα;A)\cdots\to H^k(K_\alpha,L_\alpha;A)\xrightarrow{J_k^*}H^k(K_\alpha;A)\xrightarrow{R_k^*}H^k(L_\alpha;A)\xrightarrow{\delta}H^{k+1}(K_\alpha,L_\alpha;A)\to\cdots

Proof. At cochain level, the short exact sequence

0Ck(Kα,Lα;A)JkCk(Kα;A)RkCk(Lα;A)00\to C^k(K_\alpha,L_\alpha;A)\xrightarrow{J_k}C^k(K_\alpha;A)\xrightarrow{R_k}C^k(L_\alpha;A)\to 0

holds, where Ck(Kα,Lα;A):={φCk(Kα;A):φCk(Lα)=0}C^k(K_\alpha,L_\alpha;A):=\{\varphi\in C^k(K_\alpha;A):\varphi|_{C_k(L_\alpha)}=0\}.

The coboundary operators commute with inclusion JkJ_k and restriction RkR_k: DKkJk=Jk+1DrelkD_K^k\circ J_k=J_{k+1}\circ D_{\mathrm{rel}}^k and Rk+1DKk=DLkRkR_{k+1}\circ D_K^k=D_L^k\circ R_k. The snake lemma yields the long exact sequence. \square

Theorem 12.4 (Matrix computation of relative cohomology)

For the relative coboundary matrix Drelk=Pk+1DKkIkD_{\mathrm{rel}}^k=P_{k+1}\,D_K^k\,I_k (where IkI_k injects relative basis into Ck(K)C^k(K) and Pk+1P_{k+1} projects to relative basis):

dimHk(Kα,Lα;A)=dimkerDrelkrank(Drelk1).\dim H^k(K_\alpha,L_\alpha;A)=\dim\ker D_{\mathrm{rel}}^k-\mathrm{rank}(D_{\mathrm{rel}}^{k-1}).

Proof. Rank--nullity theorem on finite-dimensional cochain complexes. \square

Lemma 12.5 (Bridge to removed-space cohomology)

Under Hypothesis 11.1 (continuous enrichment), excision and retraction give:

H(Kα,Lα;A)H(MαSα;A).H^*(K_\alpha,L_\alpha;A)\cong H^*(M_\alpha\setminus S_\alpha;A).

This separates topological assumptions from algebraic computation: the computation is performed on the finite pair (Kα,Lα)(K_\alpha,L_\alpha); the geometric interpretation is optional.

Algorithm 12.1 (Axis 2 computation)

Input: fruit complex KαK_\alpha, door subcomplex LαL_\alpha.

Output: dimHk(Kα,Lα;A)\dim H^k(K_\alpha,L_\alpha;A) for each kk.

  1. Build simplex bases for KαK_\alpha and LαL_\alpha by degree.
  2. Construct boundary matrices BkKB_k^K, BkLB_k^L and coboundary matrices DKkD_K^k, DLkD_L^k.
  3. Build relative coboundary: Drelk=Pk+1DKkIkD_{\mathrm{rel}}^k=P_{k+1}\,D_K^k\,I_k.
  4. Compute dimHk\dim H^k by rank/nullity.
  5. Verify cochain SES: RkJk=0R_k J_k=0, im(Jk)=ker(Rk)\mathrm{im}(J_k)=\ker(R_k), RkR_k surjective.
  6. Connecting map: for [αL]Hk(L)[\alpha_L]\in H^k(L), extend to α~Ck(K)\tilde\alpha\in C^k(K) with Rkα~=αLR_k\tilde\alpha=\alpha_L; then δ([αL])=[DKkα~]Hk+1(K,L)\delta([\alpha_L])=[D_K^k\tilde\alpha]\in H^{k+1}(K,L).

12.4 Axis 3: Door Set Topology + Persistence Barcode

Definition 12.3 (Persistence barcode)

The barcode of the door set Σ\Sigma is the persistence diagram of the energy-filtered simplicial complex:

Barcode(Σ):={[ai,bi)R0}\mathrm{Barcode}(\Sigma):=\{[a_i,b_i)\subset\mathbb{R}_{\ge0}\}

where aia_i is the birth energy (when a new cohomological feature appears) and bib_i is the death energy.

Construction

  1. Order doors by energy: ep1ep2epLe_{p_1}\le e_{p_2}\le\cdots\le e_{p_L}.
  2. Build filtration: L0L1LL=LαL_0\subset L_1\subset\cdots\subset L_L=L_\alpha by adding door nodes in energy order.
  3. Compute persistent cohomology across the filtration.

Theorem 12.5 (Persistence--cohomology correspondence)

dimHk(Lα;A)={i:bar [ai,bi) is a k-degree feature with bi=}.\dim H^k(L_\alpha;A)=|\{i:\text{bar }[a_i,b_i)\text{ is a }k\text{-degree feature with }b_i=\infty\}|.

Proof sketch. Standard result in persistent cohomology: the number of infinite bars in degree kk equals the Betti number βk(Lα)\beta_k(L_\alpha). \square

Definition 12.6 (Hausdorff dimension of the barcode)

dimH(Σ):=inf{δ>0:iiδ<}\dim_H(\Sigma):=\inf\Bigl\{\delta>0:\sum_i\ell_i^\delta<\infty\Bigr\}

where i=biai\ell_i=b_i-a_i is the lifespan of bar ii.

  • dimH=0\dim_H=0: isolated doors (finitely many, finite barcode).
  • 0<dimH<10<\dim_H<1: fractal door structure.
  • dimH=1\dim_H=1: maximally complex.

Theorem 12.7 (Finite door set, U(1)U(1))

If Σ={p1,,pn}\Sigma=\{p_1,\ldots,p_n\} is finite and A=U(1)A=U(1):

Hˇ0(Σ;U(1))=U(1)n,Hˇk>0(Σ;U(1))=0.\check H^0(\Sigma;U(1))=U(1)^n,\qquad\check H^{k>0}(\Sigma;U(1))=0.

Proof. A finite discrete set has trivial higher cohomology. \square


12.5 Integration: The Three-Axis Exact Sequence

Theorem 12.8 (Three-axis exact sequence)

For each degree kk, the cochain short exact sequence

0Ck(Kα,Lα;A)JkCk(Kα;A)RkCk(Lα;A)00\to C^k(K_\alpha,L_\alpha;A)\xrightarrow{J_k}C^k(K_\alpha;A)\xrightarrow{R_k}C^k(L_\alpha;A)\to 0

induces the long exact sequence in cohomology:

Hk(Kα,Lα)Axis 2Hk(Kα)Axis 1Hk(Lα)Axis 3δHk+1(Kα,Lα)Axis 2\cdots\to\underbrace{H^k(K_\alpha,L_\alpha)}_{\text{Axis 2}}\to\underbrace{H^k(K_\alpha)}_{\text{Axis 1}}\to\underbrace{H^k(L_\alpha)}_{\text{Axis 3}}\xrightarrow{\delta}\underbrace{H^{k+1}(K_\alpha,L_\alpha)}_{\text{Axis 2}}\to\cdots

This is the rigorous relationship among the three axes. \square

Interpretation: The three axes are not independent; they are linked by the connecting homomorphism δ\delta. Knowing any two determines constraints on the third.


12.6 Worked Example: 5-Node Graph, U(1)U(1)

(Continuing from Chapter 10, Section 10.2.)

Fruit F={4,5}F=\{4,5\}, door Σ={4}\Sigma=\{4\}, kernel F={5}F^\circ=\{5\}.

Fruit complex KαK_\alpha: edge (4,5)(4,5) gives a 1-simplex. Kα={[4],[5],[4,5]}K_\alpha=\{[4],[5],[4,5]\}.

Door subcomplex LαL_\alpha: vertex {4}\{4\}. Lα={[4]}L_\alpha=\{[4]\}.

Axis 1: H0(Kα;Z)=ZH^0(K_\alpha;\mathbb{Z})=\mathbb{Z} (connected), Hk>0=0H^{k>0}=0 (contractible).

Axis 2: C0(K,L)=ZC^0(K,L)=\mathbb{Z} (generated by [5][5]), C1(K,L)=ZC^1(K,L)=\mathbb{Z} (generated by [4,5][4,5]). Relative coboundary Drel0:[5][4,5]D_{\mathrm{rel}}^0:[5]\mapsto[4,5] is an isomorphism. So H0(K,L)=0H^0(K,L)=0, H1(K,L)=0H^1(K,L)=0.

Axis 3: H0(L;Z)=ZH^0(L;\mathbb{Z})=\mathbb{Z} (single point), Hk>0=0H^{k>0}=0.

Barcode: single bar [0,e4)=[0,1)[0,e_4)=[0,1), dimH=0\dim_H=0.

Exact sequence check (k=0k=0):

0H0(K,L)H0(K)H0(L)δH1(K,L)0\to H^0(K,L)\to H^0(K)\to H^0(L)\xrightarrow{\delta}H^1(K,L)\to\cdots 00ZZ00\to 0\to\mathbb{Z}\xrightarrow{\sim}\mathbb{Z}\to 0

Consistent. The map H0(K)H0(L)H^0(K)\to H^0(L) is the restriction (isomorphism since KK is connected and LL is a single point). \checkmark


12.7 Computation Protocol (Meta-Algorithm)

function ComputeWorldAxes(F, Sigma, [A_infty], A):
    // Build complexes
    K = CliqueComplex(InducedGraph(F))
    L = DoorSubcomplex(K, Sigma)
 
    // Axis 1
    H1 = SimplicialCohomology(K, A)
 
    // Axis 2
    H2 = RelativeCohomology(K, L, A)
 
    // Axis 3
    barcode = PersistentCohomology(L, energy_filtration, A)
    H3 = CohomologyFromBarcode(barcode)
    dim_H = HausdorffDimension(barcode)
 
    // Verify exact sequence
    assert LongExactConsistency(H1, H2, H3, delta)
 
    return (H1, H2, H3, barcode, dim_H)

Complexity: O(F3)O(|F|^3) (dominated by matrix rank computations on the simplicial pair).