This paper investigates the concept of vertex capacity for infinite graphs over non-Archimedean ordered fields. Unlike graphs over the real numbers, monotone limits do not necessarily exist in non-Archimedean fields. Consequently, in addition to positive and zero capacity, there exists a third case—divergent capacity. However, the authors prove that any of these cases is independent of the choice of vertex and thus constitutes a global property for connected graphs. Capacity is shown to connect energy minimization, solutions to the Dirichlet problem, and the existence of Green functions. Furthermore, the authors provide sufficient criteria in the form of a Nash-Williams test, investigate the relationship with Hardy inequalities, and discuss the existence of positive superharmonic functions. Finally, they study analytic properties in the relationship between transfer operators and the inverse of the Laplacian.
An ordered field is defined to be non-Archimedean if there exists an infinitesimal element such that for all :
A graph is defined as a triple , where:
For a finite connected subset and , the Dirichlet problem is:
\Delta v = 0 & \text{on } K\setminus\{a\} \\ v = 1 & \text{on } \{a\} \\ v = 0 & \text{on } V\setminus K \end{cases}$$ The effective capacity is defined as: $$\text{cap}_K(a) = \Delta v(a)m(a)$$ #### Capacity of Infinite Graphs Let $B_n(a) = \{x ∈ V | d(a,x) < n\}$ be the distance ball, define: $$\text{cap}_n(a) := \text{cap}_{B_n(a)}(a)$$ If the limit exists, then: $$\text{cap}(a) := \lim_{n→∞} \text{cap}_n(a)$$ ### Three Characterizations of Capacity **Theorem 4.1 (Energy minimization characterization)**: $$\text{cap}(a) = \inf\{Q(φ) | φ ∈ C_c(V), φ(a) = 1\}$$ **Theorem 4.11 (Green function characterization)**: $$\text{cap}(a) = \lim_{n→∞} \frac{m(a)}{G_n(a,a)}$$ where $G_n(x,y) = \Delta_{K_n}^{-1} 1_y(x)$. ## Main Theoretical Results ### Vertex Independence Theorem **Theorem 2.9**: If $\text{cap}(a)$ exists for some $a ∈ V$, then it exists for all $x ∈ V$. Moreover, if $\text{cap}(a) = 0$ for some $a ∈ V$, then $\text{cap}(x) = 0$ for all $x ∈ V$. ### Nash-Williams Test **Theorem 5.3**: The following conditions are equivalent: 1. $\sum_{r=0}^∞ \frac{1}{b(∂B_r)} = ∞$ 2. There exists a subsequence $(B_{n_k}(a))$ such that $\lim_{k→∞} b(∂B_{n_k}(a)) = 0$ 3. There exists a subsequence $(B_{n_k}(a))$ such that $\lim_{k→∞} \max_{(x,y)∈∂B_{n_k}(a)} b(x,y) = 0$ If any of the above conditions holds, the graph has zero capacity. ### Hardy Inequality **Theorem 6.2**: The following are equivalent: 1. The graph does not have zero capacity 2. Hardy inequality holds for some strictly positive Hardy weight 3. Hardy inequality holds for some Hardy weight ## Concrete Examples ### Weakly Spherically Symmetric Graphs For weakly spherically symmetric graphs, there is an explicit formula: $$\text{cap}(o) = \left(\sum_{k=0}^∞ \frac{1}{b(∂B_k(o))}\right)^{-1}$$ ### Path Graphs over Levi-Civita Fields - **Zero capacity example**: $b(k,k+1) = ε^k$, then $\text{cap}(0) = 0$ - **Positive capacity example**: $b(k,k+1) = ε^{-k}$, then $\text{cap}(0) = 1-ε$ - **Divergent capacity example**: $b(k,k+1) = 1$, then $\text{cap}_n(0) = 1/n$ does not converge ## Transfer Operator Analysis ### Basic Properties Define transition probability: $p(x,y) = \frac{b(x,y)}{b(x)}$ Transfer operator: $Pf(x) = \sum_{y∈V} p(x,y)f(y)$ **Lemma 8.1**: For non-Archimedean fields, the behavior of $P^n(x,y)$ is completely determined by the most probable paths. ### Inverse of the Laplacian **Theorem 8.2**: If $\lim_{n→∞} P^n(x,y) = 0$, then: $$\Delta^{-1}_K φ = \sum_{n=0}^∞ P^n_K φ$$ ## Related Work This paper builds upon the following research: 1. Capacity theory for graphs over real numbers (classical results) 2. Theory of finite graphs over non-Archimedean fields (prior work by Muranova) 3. Impedance concepts in circuit network theory 4. Random walk theory in probability ## Conclusions and Discussion ### Main Conclusions 1. Graph capacity over non-Archimedean fields exhibits three types: positive capacity, zero capacity, and divergent capacity 2. Capacity type is a global property of the graph, independent of vertex choice 3. Establishing equivalence between capacity and energy minimization, Dirichlet problems, and Green functions 4. Providing effective criteria for determining capacity types ### Limitations 1. Probabilistic interpretation in non-Archimedean fields is more subtle and requires further investigation 2. The existence of positive superharmonic functions in the divergent capacity case remains open 3. Some results require special properties of the field (such as real closure) ### Future Directions 1. Developing probability theory over non-Archimedean fields 2. Investigating more general weight function cases 3. Exploring connections with other mathematical branches ## In-Depth Evaluation ### Strengths 1. **Theoretical innovation**: First systematic establishment of complete capacity theory for graphs over non-Archimedean fields 2. **Rigorous methodology**: Clever proof techniques, particularly the proof of vertex independence 3. **Rich examples**: Concrete examples over Levi-Civita fields effectively illustrate the theory 4. **Broad applications**: Connecting analysis, algebra, probability, and other mathematical branches ### Weaknesses 1. **High technical threshold**: Requires readers to have background knowledge in non-Archimedean fields 2. **Limited practical applications**: While theoretically complete, practical application scenarios are relatively limited 3. **Open problems**: Some important questions remain unresolved (such as superharmonic functions under divergent capacity) ### Impact This work establishes important foundations for the interdisciplinary field of non-Archimedean analysis and graph theory, and is expected to promote further development in related areas, particularly in circuit network theory and nonstandard analysis. ### Applicable Scenarios 1. Circuit network analysis (particularly involving rational function weights) 2. Graph-theoretic problems in nonstandard analysis 3. Discretization problems in algebraic geometry 4. Lattice models in mathematical physics ## References The paper cites 28 important references covering classical and recent results in potential theory, graph theory, non-Archimedean analysis, and other fields, providing a solid theoretical foundation for the research.