Abstract:We study the existence of solutions of Ricci flow equations of Ollivier-Lin-Lu-Yau curvature defined on weighted graphs. Our work is motivated by\cite{NLLG} in which the discrete time Ricci flow algorithm has been applied successfully as a discrete geometric approach in detecting complex networks. Our main result is the existence and uniqueness theorem for solutions to a continuous time normalized Ricci flow. We also display possible solutions to the Ricci flow on path graph and prove the Ricci flow on finite star graph with at least three leaves converges to constant-weighted star.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is the existence and uniqueness of solutions to the Ollivier - Lin - Lu - Yau discrete Ricci flow equations on weighted graphs. Specifically, the authors focus on the existence and uniqueness of continuous - time normalized solutions to the Ricci flow equations under the Ollivier - Lin - Lu - Yau curvature defined on weighted graphs, and explore the convergence properties of these solutions.
### Background of the Paper
Ricci flow was originally proposed by Richard Hamilton in the 1980s for studying the geometric deformation of Riemannian manifolds. On a Riemannian manifold \((M, g)\), Ricci flow changes the metric \(g\) through a second - order nonlinear partial differential equation:
\[ \frac{\partial g_{ij}}{\partial t}=- 2R_{ij}, \]
where \(R_{ij}\) is the Ricci curvature. Ricci flow can smooth the metric, but may lead to the emergence of singularities, which can be removed by the surgery method. Perelman used Ricci flow and the surgery method to solve the Poincaré conjecture, and Simon Brendle and Richard Schoen used it to prove the differential sphere theorem.
### Applications of Discrete Ricci Flow
A 2019 work [13] applied Ricci flow to community detection on weighted graphs and achieved good results. The algorithm is based on Ollivier Ricci curvature and adjusts the edge weights through an iterative process so that high - weight edges are identified, thereby achieving community partitioning of the network. However, although the application of this work is excellent, it lacks rigorous results and theorems in terms of mathematical theory.
### Main Contributions of This Paper
The goal of this paper is to provide a mathematical framework for discrete Ricci flow to strictly answer the following basic questions:
1. What are the intrinsic metrics and curvatures in the graph?
2. Do solutions to the Ricci flow equations exist? What are the existence domains?
3. Do the limit objects of Ricci flow exist? Do they have constant curvature?
### Main Results
- **Existence and Uniqueness**: This paper proves that on a connected weighted graph, the solution to the normalized Ricci flow equation (4) exists and is unique in the long - term. This result also applies to the non - normalized Ricci flow equation (2).
- **Convergence**: This paper also proves the convergence results of Ricci flow on several path graphs and star graphs. For example, a path graph of length 2 will converge to a specific structure under Ricci flow, and a finite star graph will converge to a star graph with constant weights under Ricci flow.
### Mathematical Framework
- **Ollivier - Lin - Lu - Yau Curvature**: On a weighted graph \(G=(V, E, w)\), the Ollivier - Lin - Lu - Yau curvature is defined based on the optimal transport theory. Specifically, for two vertices \(x\) and \(y\), the Wasserstein distance \(W_1(\mu_x,\mu_y)\) of their probability distributions \(\mu_x\) and \(\mu_y\) is used to define the Ollivier curvature:
\[ \kappa_{xy}=1-\frac{W_1(\mu_x,\mu_y)}{d(x,y)}. \]
- **Normalized Ricci Flow Equation**: To prevent the total length of the graph edges from becoming infinitesimally small in the limit case, this paper defines the normalized Ricci flow equation:
\[ \frac{d w_e(t)}{dt}=-\kappa_e(t) w_e(t)+w_e(t)\sum_{h\in E(G)}\kappa_h(t) w_h(t). \]
### Conclusion
Through rigorous mathematical proofs, this paper provides a solid theoretical basis for the existence, uniqueness, and convergence of discrete Ricci flow, laying the foundation for future applications in complex network analysis.