Prerequisites : Chapters 2--8 (Part I).
Gauge group: G = U ( 1 ) = { e i α : α ∈ [ 0 , 2 π ) } G=U(1)=\{e^{i\alpha}:\alpha\in[0,2\pi)\} G = U ( 1 ) = { e i α : α ∈ [ 0 , 2 π )} .
Relation transit: g t ( i , j ) = e i α t ( i , j ) g_t(i,j)=e^{i\alpha_t(i,j)} g t ( i , j ) = e i α t ( i , j ) .
Gauge transformation: h ( i ) = e i φ i h(i)=e^{i\varphi_i} h ( i ) = e i φ i .
g t h ( i , j ) = e i ( α t ( i , j ) + φ i − φ j ) . g_t^h(i,j)=e^{i(\alpha_t(i,j)+\varphi_i-\varphi_j)}. g t h ( i , j ) = e i ( α t ( i , j ) + φ i − φ j ) .
For triangle ( i , j , k ) (i,j,k) ( i , j , k ) :
Ω t ( i , j , k ) = e i ( α i j + α j k + α k i ) . \Omega_t(i,j,k)=e^{i(\alpha_{ij}+\alpha_{jk}+\alpha_{ki})}. Ω t ( i , j , k ) = e i ( α ij + α j k + α k i ) .
Discrete curvature:
ω t ( i , j , k ) = 2 ( 1 − cos ( α i j + α j k + α k i ) ) . \omega_t(i,j,k)=2\bigl(1-\cos(\alpha_{ij}+\alpha_{jk}+\alpha_{ki})\bigr). ω t ( i , j , k ) = 2 ( 1 − cos ( α ij + α j k + α k i ) ) .
Since U ( 1 ) U(1) U ( 1 ) is abelian, the holonomy itself (not just its conjugacy class) is gauge-invariant.
E F ∘ ( φ ) = ∑ i , j ∈ F ∘ W t ( i , j ) > 0 W t ( i , j ) ⋅ 2 ( 1 − cos ( α i j + φ i − φ j ) ) . \mathcal{E}_{F^\circ}(\varphi)=\sum_{\substack{i,j\in F^\circ\\W_t(i,j)>0}}W_t(i,j)\cdot 2\bigl(1-\cos(\alpha_{ij}+\varphi_i-\varphi_j)\bigr). E F ∘ ( φ ) = i , j ∈ F ∘ W t ( i , j ) > 0 ∑ W t ( i , j ) ⋅ 2 ( 1 − cos ( α ij + φ i − φ j ) ) .
Critical-point equation: ∑ j ∼ i W t ( i , j ) sin ( α i j + φ i − φ j ) = 0 \sum_{j\sim i}W_t(i,j)\sin(\alpha_{ij}+\varphi_i-\varphi_j)=0 ∑ j ∼ i W t ( i , j ) sin ( α ij + φ i − φ j ) = 0 for all i ∈ F ∘ i\in F^\circ i ∈ F ∘ .
At a global minimum with small residual curvature (∣ α ~ i j ∣ < π |\tilde\alpha_{ij}|<\pi ∣ α ~ ij ∣ < π for all edges), the Hessian is a weighted graph Laplacian with positive weights (Theorem 7.11), so the minimum is unique up to a constant shift φ ↦ φ + c \varphi\mapsto\varphi+c φ ↦ φ + c .
In the small-curvature regime, the normal equations give:
L φ ∗ = − B d i a g ( W ) α , φ ∗ = − L † B d i a g ( W ) α + c ⋅ 1 L\,\varphi^*=-B\,\mathrm{diag}(W)\,\alpha,\qquad\varphi^*=-L^\dagger B\,\mathrm{diag}(W)\,\alpha+c\cdot\mathbf{1} L φ ∗ = − B diag ( W ) α , φ ∗ = − L † B diag ( W ) α + c ⋅ 1
where L = B d i a g ( W ) B T L=B\,\mathrm{diag}(W)B^T L = B diag ( W ) B T is the weighted graph Laplacian and L † L^\dagger L † its pseudoinverse.
α ~ i j = α i j + φ i ∗ − φ j ∗ , ρ ( i ) = ∑ j ∼ i W t ( i , j ) α ~ i j 2 . \tilde\alpha_{ij}=\alpha_{ij}+\varphi_i^*-\varphi_j^*,\qquad\rho(i)=\sum_{j\sim i}W_t(i,j)\,\tilde\alpha_{ij}^2. α ~ ij = α ij + φ i ∗ − φ j ∗ , ρ ( i ) = j ∼ i ∑ W t ( i , j ) α ~ ij 2 .
The constant c c c cancels in α ~ i j \tilde\alpha_{ij} α ~ ij , confirming Conjecture 7.10 for U ( 1 ) U(1) U ( 1 ) .
1 ─── 2
│ ╲ ╱ │
│ 3 │
│ ╱ ╲ │
4 ─── 5 copy
V = { 1 , 2 , 3 , 4 , 5 } V=\{1,2,3,4,5\} V = { 1 , 2 , 3 , 4 , 5 } , G = U ( 1 ) G=U(1) G = U ( 1 ) , θ = 0.15 \theta=0.15 θ = 0.15 , τ = 0.8 \tau=0.8 τ = 0.8 .
Weights:
Internal to { 1 , 2 , 3 } \{1,2,3\} { 1 , 2 , 3 } : W ( 1 , 2 ) = W ( 1 , 3 ) = W ( 2 , 3 ) = 10 W(1,2)=W(1,3)=W(2,3)=10 W ( 1 , 2 ) = W ( 1 , 3 ) = W ( 2 , 3 ) = 10 .
Internal to { 4 , 5 } \{4,5\} { 4 , 5 } : W ( 4 , 5 ) = 8 W(4,5)=8 W ( 4 , 5 ) = 8 .
Bridge: W ( 3 , 4 ) = 1 W(3,4)=1 W ( 3 , 4 ) = 1 , W ( 2 , 5 ) = 0.5 W(2,5)=0.5 W ( 2 , 5 ) = 0.5 .
All others: 0 0 0 .
i i i d ( i ) d(i) d ( i ) Description 1 10 + 10 = 20 10+10=20 10 + 10 = 20 Deep interior of { 1 , 2 , 3 } \{1,2,3\} { 1 , 2 , 3 } 2 10 + 10 + 0.5 = 20.5 10+10+0.5=20.5 10 + 10 + 0.5 = 20.5 Boundary (connected to 5) 3 10 + 10 + 1 = 21 10+10+1=21 10 + 10 + 1 = 21 Boundary (connected to 4) 4 8 + 1 = 9 8+1=9 8 + 1 = 9 Boundary of { 4 , 5 } \{4,5\} { 4 , 5 } 5 8 + 0.5 = 8.5 8+0.5=8.5 8 + 0.5 = 8.5 Boundary of { 4 , 5 } \{4,5\} { 4 , 5 }
Candidate F = { 4 , 5 } F=\{4,5\} F = { 4 , 5 } :
v o l ( { 4 , 5 } ) = 9 + 8.5 = 17.5 \mathrm{vol}(\{4,5\})=9+8.5=17.5 vol ({ 4 , 5 }) = 9 + 8.5 = 17.5 .
v o l ( V ) = 20 + 20.5 + 21 + 9 + 8.5 = 79 \mathrm{vol}(V)=20+20.5+21+9+8.5=79 vol ( V ) = 20 + 20.5 + 21 + 9 + 8.5 = 79 .
(F1): 17.5 ≤ 39.5 = 1 2 ⋅ 79 17.5\le 39.5=\frac{1}{2}\cdot 79 17.5 ≤ 39.5 = 2 1 ⋅ 79 . Yes.
c u t ( { 4 , 5 } , { 1 , 2 , 3 } ) = W ( 4 , 3 ) + W ( 5 , 2 ) = 1 + 0.5 = 1.5 \mathrm{cut}(\{4,5\},\{1,2,3\})=W(4,3)+W(5,2)=1+0.5=1.5 cut ({ 4 , 5 } , { 1 , 2 , 3 }) = W ( 4 , 3 ) + W ( 5 , 2 ) = 1 + 0.5 = 1.5 .
ϕ ( { 4 , 5 } ) = 1.5 / 17.5 ≈ 0.086 ≤ 0.15 \phi(\{4,5\})=1.5/17.5\approx 0.086\le 0.15 ϕ ({ 4 , 5 }) = 1.5/17.5 ≈ 0.086 ≤ 0.15 . Yes.
{ 4 , 5 } ∈ F t \{4,5\}\in\mathfrak{F}_t { 4 , 5 } ∈ F t .
Candidate F = { 1 , 2 , 3 } F=\{1,2,3\} F = { 1 , 2 , 3 } : v o l ( { 1 , 2 , 3 } ) = 61.5 > 39.5 \mathrm{vol}(\{1,2,3\})=61.5>39.5 vol ({ 1 , 2 , 3 }) = 61.5 > 39.5 . Violates (F1). Not a fruit.
For F = { 4 , 5 } F=\{4,5\} F = { 4 , 5 } :
i i i d ( i ) d(i) d ( i ) ∑ j ∈ F W ( i , j ) \sum_{j\in F}W(i,j) ∑ j ∈ F W ( i , j ) b F , t ( i ) b_{F,t}(i) b F , t ( i ) Door? 4 9 8 1 1 ≥ 0.8 1\ge 0.8 1 ≥ 0.8 : Yes 5 8.5 8 0.5 0.5 < 0.8 0.5<0.8 0.5 < 0.8 : No
Σ = { 4 } \Sigma=\{4\} Σ = { 4 } , e 4 = 1 e_4=1 e 4 = 1 .
F ∘ = { 4 , 5 } ∖ { 4 } = { 5 } F^\circ=\{4,5\}\setminus\{4\}=\{5\} F ∘ = { 4 , 5 } ∖ { 4 } = { 5 } . Single node, no internal edges.
[ A ∞ ] [A_\infty] [ A ∞ ] is trivial (empty connection). E x i s t e n c e ( { 4 , 5 } , t ) = ( [ ∅ ] , { 4 } , { 1 } ) \mathrm{Existence}(\{4,5\},t)=([\emptyset],\{4\},\{1\}) Existence ({ 4 , 5 } , t ) = ([ ∅ ] , { 4 } , { 1 }) .
Theorem A : Internal energy = 2 ⋅ 8 = 16 =2\cdot 8=16 = 2 ⋅ 8 = 16 ; v o l = 17.5 \mathrm{vol}=17.5 vol = 17.5 ; ratio = 16 / 17.5 ≈ 0.914 ≥ 1 − 0.15 = 0.85 =16/17.5\approx 0.914\ge 1-0.15=0.85 = 16/17.5 ≈ 0.914 ≥ 1 − 0.15 = 0.85 . Satisfied.
Theorem B : ∣ Σ ∣ = 1 ≤ 0.15 ⋅ 17.5 / 0.8 ≈ 3.28 |\Sigma|=1\le 0.15\cdot 17.5/0.8\approx 3.28 ∣Σ∣ = 1 ≤ 0.15 ⋅ 17.5/0.8 ≈ 3.28 . Satisfied.
Theorem D : Expected escape time ≥ 1 / ( 2 ⋅ 0.15 ) ≈ 3.3 \ge 1/(2\cdot 0.15)\approx 3.3 ≥ 1/ ( 2 ⋅ 0.15 ) ≈ 3.3 steps. Reasonable for a 2-node fruit.
Fruit A: {1,2,3,4} — complete graph, internal weight 10
Fruit B: {7,8,9,10} — complete graph, internal weight 8
Stem: {5,6}
Bridges:
3 -- 5 (weight 0.5)
4 -- 5 (weight 0.3)
5 -- 6 (weight 2.0)
6 -- 7 (weight 0.4)
6 -- 8 (weight 0.6) copy
Parameters: θ = 0.1 \theta=0.1 θ = 0.1 , τ = 0.5 \tau=0.5 τ = 0.5 .
Fruit A internal edges (complete K 4 K_4 K 4 , weight 10): each node has 3 internal edges, so internal degree = 3 × 10 = 30 =3\times 10=30 = 3 × 10 = 30 .
Node Internal deg Bridge deg d ( i ) d(i) d ( i ) 1 30 0 30 2 30 0 30 3 30 0.5 30.5 4 30 0.3 30.3 5 0 0.5 + 0.3 + 2.0 = 2.8 0.5+0.3+2.0=2.8 0.5 + 0.3 + 2.0 = 2.8 2.8 6 0 2.0 + 0.4 + 0.6 = 3.0 2.0+0.4+0.6=3.0 2.0 + 0.4 + 0.6 = 3.0 3.0 7 3 × 8 = 24 3\times 8=24 3 × 8 = 24 0.4 24.4 8 24 0.6 24.6 9 24 0 24 10 24 0 24
v o l ( V ) = 30 + 30 + 30.5 + 30.3 + 2.8 + 3.0 + 24.4 + 24.6 + 24 + 24 = 223.6 \mathrm{vol}(V)=30+30+30.5+30.3+2.8+3.0+24.4+24.6+24+24=223.6 vol ( V ) = 30 + 30 + 30.5 + 30.3 + 2.8 + 3.0 + 24.4 + 24.6 + 24 + 24 = 223.6 .
Fruit A = { 1 , 2 , 3 , 4 } =\{1,2,3,4\} = { 1 , 2 , 3 , 4 } :
v o l ( A ) = 30 + 30 + 30.5 + 30.3 = 120.8 \mathrm{vol}(A)=30+30+30.5+30.3=120.8 vol ( A ) = 30 + 30 + 30.5 + 30.3 = 120.8 .
(F1): 120.8 > 111.8 = 1 2 ⋅ 223.6 120.8>111.8=\frac{1}{2}\cdot 223.6 120.8 > 111.8 = 2 1 ⋅ 223.6 . Violates (F1) .
So { 1 , 2 , 3 , 4 } \{1,2,3,4\} { 1 , 2 , 3 , 4 } is too large (majority). Check complement:
Fruit B = { 7 , 8 , 9 , 10 } =\{7,8,9,10\} = { 7 , 8 , 9 , 10 } :
v o l ( B ) = 24.4 + 24.6 + 24 + 24 = 97.0 \mathrm{vol}(B)=24.4+24.6+24+24=97.0 vol ( B ) = 24.4 + 24.6 + 24 + 24 = 97.0 .
(F1): 97.0 ≤ 111.8 97.0\le 111.8 97.0 ≤ 111.8 . Yes.
c u t ( B , B ˉ ) = W ( 7 , 6 ) + W ( 8 , 6 ) = 0.4 + 0.6 = 1.0 \mathrm{cut}(B,\bar B)=W(7,6)+W(8,6)=0.4+0.6=1.0 cut ( B , B ˉ ) = W ( 7 , 6 ) + W ( 8 , 6 ) = 0.4 + 0.6 = 1.0 .
ϕ ( B ) = 1.0 / 97.0 ≈ 0.0103 ≤ 0.1 \phi(B)=1.0/97.0\approx 0.0103\le 0.1 ϕ ( B ) = 1.0/97.0 ≈ 0.0103 ≤ 0.1 . Yes.
{ 7 , 8 , 9 , 10 } ∈ F t \{7,8,9,10\}\in\mathfrak{F}_t { 7 , 8 , 9 , 10 } ∈ F t .
Checking smaller subsets : { 5 , 6 } \{5,6\} { 5 , 6 } has v o l = 5.8 \mathrm{vol}=5.8 vol = 5.8 , c u t = 0.5 + 0.3 + 0.4 + 0.6 = 1.8 \mathrm{cut}=0.5+0.3+0.4+0.6=1.8 cut = 0.5 + 0.3 + 0.4 + 0.6 = 1.8 , ϕ = 1.8 / 5.8 ≈ 0.31 > 0.1 \phi=1.8/5.8\approx 0.31>0.1 ϕ = 1.8/5.8 ≈ 0.31 > 0.1 . Not a fruit.
Summary of fruits : F t = { { 7 , 8 , 9 , 10 } } \mathfrak{F}_t=\{\{7,8,9,10\}\} F t = {{ 7 , 8 , 9 , 10 }} (only Fruit B qualifies under the strict (F1) condition).
Remark. If we relax (F1) or embed this graph in a larger graph so that v o l ( A ) ≤ 1 2 v o l ( V ) \mathrm{vol}(A)\le\frac{1}{2}\mathrm{vol}(V) vol ( A ) ≤ 2 1 vol ( V ) , then A A A would also be a fruit with ϕ ≈ 0.007 \phi\approx 0.007 ϕ ≈ 0.007 .
Node d ( i ) d(i) d ( i ) ∑ j ∈ B W ( i , j ) \sum_{j\in B}W(i,j) ∑ j ∈ B W ( i , j ) b B , t ( i ) b_{B,t}(i) b B , t ( i ) Door? 7 24.4 24 0.4 0.4 < 0.5 0.4<0.5 0.4 < 0.5 : No8 24.6 24 0.6 0.6 ≥ 0.5 0.6\ge 0.5 0.6 ≥ 0.5 : Yes 9 24 24 0 No 10 24 24 0 No
Σ = { 8 } \Sigma=\{8\} Σ = { 8 } , e 8 = 0.6 e_8=0.6 e 8 = 0.6 .
B ∘ = { 7 , 9 , 10 } B^\circ=\{7,9,10\} B ∘ = { 7 , 9 , 10 } .
Internal edges of B ∘ B^\circ B ∘ : ( 7 , 9 ) , ( 7 , 10 ) , ( 9 , 10 ) (7,9),(7,10),(9,10) ( 7 , 9 ) , ( 7 , 10 ) , ( 9 , 10 ) , all with weight 8. This is a complete triangle K 3 K_3 K 3 .
Transit angles (assign): α 79 = 0.3 \alpha_{79}=0.3 α 79 = 0.3 , α 7 , 10 = 0.1 \alpha_{7,10}=0.1 α 7 , 10 = 0.1 , α 9 , 10 = 0.2 \alpha_{9,10}=0.2 α 9 , 10 = 0.2 .
Holonomy of triangle ( 7 , 9 , 10 ) (7,9,10) ( 7 , 9 , 10 ) : Ω = e i ( 0.3 + 0.2 − 0.1 ) = e 0.4 i \Omega=e^{i(0.3+0.2-0.1)}=e^{0.4i} Ω = e i ( 0.3 + 0.2 − 0.1 ) = e 0.4 i . Non-trivial curvature.
Optimal gauge (small-curvature approximation):
Incidence matrix for K 3 K_3 K 3 on { 7 , 9 , 10 } \{7,9,10\} { 7 , 9 , 10 } (edges ordered as e 1 = ( 7 , 9 ) , e 2 = ( 7 , 10 ) , e 3 = ( 9 , 10 ) e_1=(7,9),e_2=(7,10),e_3=(9,10) e 1 = ( 7 , 9 ) , e 2 = ( 7 , 10 ) , e 3 = ( 9 , 10 ) ):
B = ( − 1 1 0 − 1 0 1 0 − 1 1 ) , W = d i a g ( 8 , 8 , 8 ) = 8 I . B=\begin{pmatrix}-1&1&0\\-1&0&1\\0&-1&1\end{pmatrix},\quad W=\mathrm{diag}(8,8,8)=8I. B = − 1 − 1 0 1 0 − 1 0 1 1 , W = diag ( 8 , 8 , 8 ) = 8 I .
L = B W B T = 8 ( 2 − 1 − 1 − 1 2 − 1 − 1 − 1 2 ) . L=BWB^T=8\begin{pmatrix}2&-1&-1\\-1&2&-1\\-1&-1&2\end{pmatrix}. L = B W B T = 8 2 − 1 − 1 − 1 2 − 1 − 1 − 1 2 .
α = ( 0.3 , 0.1 , 0.2 ) T \alpha=(0.3,0.1,0.2)^T α = ( 0.3 , 0.1 , 0.2 ) T .
B d i a g ( W ) α = 8 B α = 8 ( − 0.3 + 0.1 − 0.3 + 0.2 − 0.1 + 0.2 ) = 8 ( − 0.2 − 0.1 0.1 ) = ( − 1.6 − 0.8 0.8 ) . B\,\mathrm{diag}(W)\alpha=8B\alpha=8\begin{pmatrix}-0.3+0.1\\-0.3+0.2\\-0.1+0.2\end{pmatrix}=8\begin{pmatrix}-0.2\\-0.1\\0.1\end{pmatrix}=\begin{pmatrix}-1.6\\-0.8\\0.8\end{pmatrix}. B diag ( W ) α = 8 B α = 8 − 0.3 + 0.1 − 0.3 + 0.2 − 0.1 + 0.2 = 8 − 0.2 − 0.1 0.1 = − 1.6 − 0.8 0.8 .
Normal equation: L φ ∗ = ( 1.6 0.8 − 0.8 ) L\varphi^*=\begin{pmatrix}1.6\\0.8\\-0.8\end{pmatrix} L φ ∗ = 1.6 0.8 − 0.8 (note sign).
L L L has eigenvalues 0 , 24 , 24 0,24,24 0 , 24 , 24 with ker L = s p a n ( ( 1 , 1 , 1 ) T ) \ker L=\mathrm{span}((1,1,1)^T) ker L = span (( 1 , 1 , 1 ) T ) . Projecting out the kernel:
φ ∗ = L † ( 1.6 0.8 − 0.8 ) + c ( 1 1 1 ) . \varphi^*=L^\dagger\begin{pmatrix}1.6\\0.8\\-0.8\end{pmatrix}+c\begin{pmatrix}1\\1\\1\end{pmatrix}. φ ∗ = L † 1.6 0.8 − 0.8 + c 1 1 1 .
The RHS has zero mean already: 1.6 + 0.8 − 0.8 = 1.6 ≠ 0 1.6+0.8-0.8=1.6\ne 0 1.6 + 0.8 − 0.8 = 1.6 = 0 . Project:
v ˉ = 1.6 / 3 ≈ 0.533 \bar v=1.6/3\approx 0.533 v ˉ = 1.6/3 ≈ 0.533 . Centered: ( 1.067 , 0.267 , − 1.333 ) (1.067, 0.267, -1.333) ( 1.067 , 0.267 , − 1.333 ) .
L † = 1 24 ( 2 / 3 − 1 / 3 − 1 / 3 − 1 / 3 2 / 3 − 1 / 3 − 1 / 3 − 1 / 3 2 / 3 ) L^\dagger=\frac{1}{24}\begin{pmatrix}2/3&-1/3&-1/3\\-1/3&2/3&-1/3\\-1/3&-1/3&2/3\end{pmatrix} L † = 24 1 2/3 − 1/3 − 1/3 − 1/3 2/3 − 1/3 − 1/3 − 1/3 2/3 (pseudoinverse of L L L on the orthogonal complement of 1 \mathbf{1} 1 ).
Actually, for a 3 × 3 3\times 3 3 × 3 Laplacian L = 8 ⋅ 3 ( I − 1 3 11 T ) = 24 ( I − 1 3 11 T ) L=8\cdot 3(I-\frac{1}{3}\mathbf{1}\mathbf{1}^T)=24(I-\frac{1}{3}\mathbf{1}\mathbf{1}^T) L = 8 ⋅ 3 ( I − 3 1 1 1 T ) = 24 ( I − 3 1 1 1 T ) , the pseudoinverse is L † = 1 24 ( I − 1 3 11 T ) L^\dagger=\frac{1}{24}(I-\frac{1}{3}\mathbf{1}\mathbf{1}^T) L † = 24 1 ( I − 3 1 1 1 T ) .
φ ∗ = 1 24 ( I − 1 3 11 T ) ( 1.6 0.8 − 0.8 ) = 1 24 ( 1.6 − 0.533 0.8 − 0.533 − 0.8 − 0.533 ) = 1 24 ( 1.067 0.267 − 1.333 ) ≈ ( 0.0444 0.0111 − 0.0556 ) . \varphi^*=\frac{1}{24}(I-\frac{1}{3}\mathbf{1}\mathbf{1}^T)\begin{pmatrix}1.6\\0.8\\-0.8\end{pmatrix}=\frac{1}{24}\begin{pmatrix}1.6-0.533\\0.8-0.533\\-0.8-0.533\end{pmatrix}=\frac{1}{24}\begin{pmatrix}1.067\\0.267\\-1.333\end{pmatrix}\approx\begin{pmatrix}0.0444\\0.0111\\-0.0556\end{pmatrix}. φ ∗ = 24 1 ( I − 3 1 1 1 T ) 1.6 0.8 − 0.8 = 24 1 1.6 − 0.533 0.8 − 0.533 − 0.8 − 0.533 = 24 1 1.067 0.267 − 1.333 ≈ 0.0444 0.0111 − 0.0556 .
(Taking c = 0 c=0 c = 0 .)
Residual angles :
α ~ 79 = 0.3 + ( 0.0444 − 0.0111 ) = 0.3 + 0.0333 = 0.333 \tilde\alpha_{79}=0.3+(0.0444-0.0111)=0.3+0.0333=0.333 α ~ 79 = 0.3 + ( 0.0444 − 0.0111 ) = 0.3 + 0.0333 = 0.333 .
α ~ 7 , 10 = 0.1 + ( 0.0444 − ( − 0.0556 ) ) = 0.1 + 0.1 = 0.2 \tilde\alpha_{7,10}=0.1+(0.0444-(-0.0556))=0.1+0.1=0.2 α ~ 7 , 10 = 0.1 + ( 0.0444 − ( − 0.0556 )) = 0.1 + 0.1 = 0.2 .
α ~ 9 , 10 = 0.2 + ( 0.0111 − ( − 0.0556 ) ) = 0.2 + 0.0667 = 0.267 \tilde\alpha_{9,10}=0.2+(0.0111-(-0.0556))=0.2+0.0667=0.267 α ~ 9 , 10 = 0.2 + ( 0.0111 − ( − 0.0556 )) = 0.2 + 0.0667 = 0.267 .
Check : holonomy = 0.333 + 0.267 − 0.2 = 0.4 =0.333+0.267-0.2=0.4 = 0.333 + 0.267 − 0.2 = 0.4 (unchanged, as expected---gauge transformation preserves holonomy).
Residual curvature :
ρ ( 7 ) = 8 ( 0.333 2 + 0.2 2 ) = 8 ( 0.111 + 0.04 ) = 1.208 \rho(7)=8(0.333^2+0.2^2)=8(0.111+0.04)=1.208 ρ ( 7 ) = 8 ( 0.33 3 2 + 0. 2 2 ) = 8 ( 0.111 + 0.04 ) = 1.208 .
ρ ( 9 ) = 8 ( 0.333 2 + 0.267 2 ) = 8 ( 0.111 + 0.071 ) = 1.456 \rho(9)=8(0.333^2+0.267^2)=8(0.111+0.071)=1.456 ρ ( 9 ) = 8 ( 0.33 3 2 + 0.26 7 2 ) = 8 ( 0.111 + 0.071 ) = 1.456 .
ρ ( 10 ) = 8 ( 0.2 2 + 0.267 2 ) = 8 ( 0.04 + 0.071 ) = 0.888 \rho(10)=8(0.2^2+0.267^2)=8(0.04+0.071)=0.888 ρ ( 10 ) = 8 ( 0. 2 2 + 0.26 7 2 ) = 8 ( 0.04 + 0.071 ) = 0.888 .
Total energy : E = 8 ( 0.333 2 + 0.2 2 + 0.267 2 ) = 8 ( 0.111 + 0.04 + 0.071 ) = 1.776 \mathcal{E}=8(0.333^2+0.2^2+0.267^2)=8(0.111+0.04+0.071)=1.776 E = 8 ( 0.33 3 2 + 0. 2 2 + 0.26 7 2 ) = 8 ( 0.111 + 0.04 + 0.071 ) = 1.776 .
Existence of Fruit B :
E x i s t e n c e ( { 7 , 8 , 9 , 10 } , t ) = ( [ g ~ ∣ B ∘ ] G c o n s t , { 8 } , { 0.6 } ) \mathrm{Existence}(\{7,8,9,10\},t)=\bigl([\tilde g|_{B^\circ}]_{\mathcal{G}_{\mathrm{const}}},\;\{8\},\;\{0.6\}\bigr) Existence ({ 7 , 8 , 9 , 10 } , t ) = ( [ g ~ ∣ B ∘ ] G const , { 8 } , { 0.6 } )
where g ~ \tilde g g ~ encodes residual angles ( 0.333 , 0.2 , 0.267 ) (0.333, 0.2, 0.267) ( 0.333 , 0.2 , 0.267 ) on the triangle { 7 , 9 , 10 } \{7,9,10\} { 7 , 9 , 10 } .
Theorem A : Internal energy = 6 × 2 × 8 = 96 =6\times 2\times 8=96 = 6 × 2 × 8 = 96 (6 directed pairs in K 4 K_4 K 4 , each weight 8). v o l ( B ) = 97 \mathrm{vol}(B)=97 vol ( B ) = 97 . Ratio = 96 / 97 ≈ 0.990 ≥ 1 − 0.1 = 0.9 =96/97\approx 0.990\ge 1-0.1=0.9 = 96/97 ≈ 0.990 ≥ 1 − 0.1 = 0.9 . Satisfied.
Theorem B : ∣ Σ ∣ = 1 ≤ 0.1 × 97 / 0.5 = 19.4 |\Sigma|=1\le 0.1\times 97/0.5=19.4 ∣Σ∣ = 1 ≤ 0.1 × 97/0.5 = 19.4 . Satisfied.
Theorem E : Curvature is distributed over the K 3 K_3 K 3 kernel. Node 10 (farthest from door 8) has the smallest ρ \rho ρ , consistent with localisation.
G = S U ( 2 ) G=SU(2) G = S U ( 2 ) , with elements g = ( a − b ˉ b a ˉ ) g=\begin{pmatrix}a&-\bar b\\b&\bar a\end{pmatrix} g = ( a b − b ˉ a ˉ ) , ∣ a ∣ 2 + ∣ b ∣ 2 = 1 |a|^2+|b|^2=1 ∣ a ∣ 2 + ∣ b ∣ 2 = 1 .
Holonomy conjugacy class ≠ \ne = holonomy itself (base-point dependence remains).
Scalar curvature ω t ( △ ) = d S U ( 2 ) ( Ω , I ) 2 \omega_t(\triangle)=d_{SU(2)}(\Omega,I)^2 ω t ( △ ) = d S U ( 2 ) ( Ω , I ) 2 is still gauge-invariant.
Energy landscape E F ∘ ( h ) \mathcal{E}_{F^\circ}(h) E F ∘ ( h ) is non-convex: multiple local minima possible.
{ 1 , 2 , 3 } \{1,2,3\} { 1 , 2 , 3 } with:
g ( 1 , 2 ) = e i π σ x / 4 , g ( 2 , 3 ) = e i π σ y / 4 , g ( 3 , 1 ) = e i π σ z / 4 g(1,2)=e^{i\pi\sigma_x/4},\quad g(2,3)=e^{i\pi\sigma_y/4},\quad g(3,1)=e^{i\pi\sigma_z/4} g ( 1 , 2 ) = e iπ σ x /4 , g ( 2 , 3 ) = e iπ σ y /4 , g ( 3 , 1 ) = e iπ σ z /4
Holonomy: Ω = e i π σ x / 4 ⋅ e i π σ y / 4 ⋅ e i π σ z / 4 ≠ I \Omega=e^{i\pi\sigma_x/4}\cdot e^{i\pi\sigma_y/4}\cdot e^{i\pi\sigma_z/4}\ne I Ω = e iπ σ x /4 ⋅ e iπ σ y /4 ⋅ e iπ σ z /4 = I . Non-zero curvature that cannot be removed by gauge transformation (topological obstruction from holonomy).
This illustrates that for non-abelian G G G , complete flattening may be impossible, and the residual curvature reflects genuine topological content.