Rich Flow Admissible Graphs

• Rich flow admissible graphs are multigraphs that admit a nowhere-zero k-flow with the additional constraint that adjacent edges receive distinct absolute flow values.
• A recent breakthrough shows that any rich flow admissible graph with maximum degree Δ supports a rich (264Δ - 445)-flow, marking a key advance in local flow extremality.
• Structural methodologies leveraging group superposition and decomposition techniques ensure global consistency while resolving local adjacent edge conflicts.

A rich flow admissible graph is a graph that admits a nowhere-zero $k$-flow in which each pair of adjacent edges are assigned distinct absolute values. This concept is a strengthening of the classic theory of nowhere-zero flows, adding local distinctness constraints on flow magnitudes, and has led to new structural and extremal questions in flow theory. Recent advances have produced the first linear upper bound on the rich flow number as a function of maximum degree, marking a significant step toward the understanding of local extremality in graph flows (Lukoťka, 22 Jan 2026).

1. Definitions and Preliminaries

Let $G=(V,E)$ be a multigraph (no loops but multiple edges allowed), equipped with a fixed orientation of each edge.

• Nowhere-zero $k$-flow: A function $\varphi: E \rightarrow \mathbb{Z}$ is a nowhere-zero $k$-flow if
  • $0 < |\varphi(e)| < k$ for every $e \in E$,
  • and for every $v \in V$, the net flow in equals the net flow out:

  $$\sum_{e \to v} \varphi(e) = \sum_{v \to f} \varphi(f).$$

• Rich $k$-flow: A nowhere-zero $k$-flow $\varphi$ such that for all adjacent edges $e,f$, $|\varphi(e)| \neq |\varphi(f)|$.

• Rich flow number ($\phi_r(G)$): The smallest $k \in \mathbb{N}$ such that $G$ admits a rich $k$-flow.

• Rich-flow-admissible graph: $G$ is rich-flow-admissible if and only if

  • $G$ is bridgeless (contains no edge whose removal disconnects the graph), and
  • no two edges of any $2$-edge-cut share a common end-vertex.

This set of conditions is more restrictive than that for classical nowhere-zero flows, reflecting the local uniqueness demands of the rich flow property.

2. Linear Upper Bound on the Rich Flow Number

The central result is the establishment of a linear bound for the rich flow number in terms of the maximum degree $\Delta$ of a graph. Specifically:

Theorem.

If $G$ is a rich-flow-admissible graph of maximum degree $\Delta$, then

$$G \text{ admits a rich } (264\Delta - 445)\text{-flow}.$$

This result, the first of its kind, demonstrates that the rich flow number grows at most linearly with the maximum degree, paralleling (though not matching) classic results in ordinary nowhere-zero flow theory (Lukoťka, 22 Jan 2026).

Sketch of the Proof Construction

The proof constructs the required flow by superimposing flows over several abelian groups:

• Step 1: For 3-edge-connected $G$, construct a nowhere-zero $\mathbb{Z}_k \times \mathbb{Z}_2$-flow for $k = 8\Delta-13$, ensuring that adjacent-edge constraints are satisfied.
• Step 2: For graphs not 3-edge-connected, decompose along a minimal $2$-edge-cut, reassemble using auxiliary edges, and paste local flows to obtain global consistency.
• Step 3: Apply a nowhere-zero $\mathbb{Z}_6$-flow to further refine the assignment, invoking Seymour’s $6$-flow theorem to eliminate persistent adjacent conflicts.
• Step 4: The final integer-valued flow is assembled as

$\varphi = \varphi_3 + 11(\varphi_2 + 3\varphi_1),$

where $\varphi_1, \varphi_2, \varphi_3$ are the flows over $\mathbb{Z}_k, \mathbb{Z}_2, \mathbb{Z}_6$, respectively.

A direct analysis confirms that the absolute value of every edge’s flow does not exceed $264\Delta - 446$, and comprehensive case analysis eliminates any instance of adjacent edges receiving flows of the same absolute value.

3. Structural Examples and Tightness

A key example illustrating tightness involves the multigraph with three vertices $a,b,c$ joined by $k>1$ parallel edges between each pair:

• Here, $\Delta = 2k$, the edge-chromatic number is $3k$, and the rich flow number is $\phi_r(G) = 3k+1$.
• This matches the established upper bound up to an additive constant as $k$ grows, suggesting the linear bound is essentially best possible up to constant factors (Lukoťka, 22 Jan 2026).

Comparison with Tutte’s and Seymour’s classic flow results is instructive: while every bridgeless graph is conjectured by Tutte to admit a nowhere-zero $5$-flow (and is known to admit a $6$-flow by Seymour’s theorem), the rich flow requirement—adjacent absolute value distinction—strengthens these classical conditions. It follows directly that $\phi_r(G) \geq \chi'(G) + 1 \geq \Delta + 1$, where $\chi'(G)$ denotes the edge chromatic number.

4. Interplay With Broader Flow Theory

The investigation of rich flows occupies a significant niche within extremal and local graph flow problems. Whereas classical nowhere-zero flows concern global assignment constraints, the rich flow property introduces a fundamentally local extremal challenge: adjacent edges must be assigned flow values with distinct magnitudes. The transition from classical to rich flows can be seen as analogous to moving from edge coloring to vertex distinguishing edge coloring, sharply increasing structural and combinatorial complexity.

The methodology draws deeply on structural decomposition, group flows, and combinatorial circuit analysis, including chain-connect techniques, auxiliary graph transformations, and careful avoidance of confluent and contrafluent adjacency patterns, all while retaining tight control over possible forbidden residues at each step (Lukoťka, 22 Jan 2026).

5. Open Problems and Conjectures

Despite the advance represented by the $264\Delta-445$ bound, this value is not believed to be optimal. The following conjectures remain central:

• Conjecture 4.1: Every rich-flow-admissible graph of maximum degree $\Delta \geq 5$ admits a rich $\lfloor 1.5\Delta + 1 \rfloor$-flow.
• Conjecture 4.2: Every $3$-edge-connected graph of maximum degree $\Delta$ admits a rich $(\Delta + 3)$-flow.

It is further noted that if only confluent (or only contrafluent) adjacent pairs are forbidden, much sharper bounds can be established, in some cases as low as $O(\ln\Delta)$. However, the fully rich case, in which all adjacent absolute values must differ, stubbornly resists such improvements and appears to demand either more refined structural decompositions or probabilistic flow methods (Lukoťka, 22 Jan 2026).

6. Connections to Other Extremal Flow Notions

The study of rich flow admissible graphs is part of a broader movement to refine and extend classical flow theory with local distinguishing constraints—paralleling advances in distinguishing colorings, local antimagic labelings, and related combinatorial optimization problems. Recent construction techniques and group-superposition arguments developed in the context of rich flows may find further application in adjacent graph invariants and generalized flow problems, considerably enriching the theoretical landscape of graph flows.