Graph Orientation Reconfiguration Problem (GORP)

• The Graph Orientation Reconfiguration Problem (GORP) is defined as determining whether one valid graph orientation can be reached from another by sequential single-edge flips that maintain feasibility.
• It is approached using two formalisms: a homomorphism-based reconfiguration framework capturing global cyclic constraints and a local vertex constraint model akin to SAT-E2 systems.
• Algorithmic advances leverage groupoid theory and forbidden cycle structures, yielding polynomial-time solutions for specific digraph templates and orientation settings.

Searching arXiv for the cited papers and closely related work on reconfiguration and graph orientation. The Graph Orientation Reconfiguration Problem (GORP) asks whether one feasible orientation of a graph can be transformed into another by flipping one edge at a time while maintaining feasibility at every intermediate step. In the current literature represented by "Reconfiguration of Digraph Homomorphisms" (Lévêque et al., 2022) and "Planar Graph Orientation Frameworks, Applied to KPlumber and Polyomino Tiling" (Group et al., 3 Mar 2026), GORP is treated most naturally through two adjacent formalisms: reconfiguration of digraph homomorphisms, where feasible orientations are encoded as homomorphisms into a fixed digraph $H$, and graph orientation under local vertex constraints, where feasibility is determined by allowed in-degree patterns at each vertex. Together, these works place GORP within combinatorial reconfiguration, digraph homomorphism theory, and SAT-E2-style orientation constraint systems.

1. Problem formulations and equivalent viewpoints

GORP is the reconfiguration counterpart of a graph orientation feasibility problem. In the orientation language, the question is whether two feasible orientations can be connected by a sequence of single-edge flips that preserves the defining constraints throughout. The data identifies two principal ways to formalize such constraints.

The first viewpoint is homomorphism-based. Given digraphs $G$ and $H$, a homomorphism $\phi: G \to H$ maps vertices so that every arc of $G$ is mapped to an arc of $H$. The corresponding $H$-Recoloring problem asks whether two homomorphisms $\alpha,\beta: G \to H$ can be transformed into one another by changing the image of a single vertex of $G$ in each step, while maintaining a homomorphism at every step. Its reconfiguration graph $\mathcal{R}_H(G)$ has as vertices the $G$0-colorings of $G$1, with edges between colorings differing at exactly one vertex. When $G$2 is an orientation pattern or a digraph encoding global forbidden patterns, GORP reduces to an $G$3-Recoloring problem (Lévêque et al., 2022).

The second viewpoint is local-constraint graph orientation. In the Graph Orientation (GO) problem, the input is an undirected graph in which each vertex $G$4 of degree $G$5 is assigned a vertex type $G$6, specifying which in-degrees are allowed at $G$7. The question is whether the edges can be oriented so that every vertex has in-degree in its assigned set. In the symmetric setting, only the number of incoming edges matters, not which incident edges are incoming. Planar GO is the same problem restricted to planar graphs (Group et al., 3 Mar 2026).

These two formulations illuminate different aspects of GORP. The homomorphism formulation emphasizes global structural constraints and reconfiguration paths in $G$8. The local-constraint formulation emphasizes orientation feasibility as a SAT-like constraint system. A plausible implication is that a comprehensive theory of GORP must combine both perspectives rather than treating orientation reconfiguration as purely local or purely topological.

2. Homomorphism reconfiguration as an orientation framework

The digraph-homomorphism perspective provides a direct route from orientation problems to reconfiguration. "Reconfiguration of Digraph Homomorphisms" extends an algorithm of Wrochna for $G$9-Recoloring when $H$0 is a square-free loopless undirected graph to the more general setting of directed graphs (Lévêque et al., 2022). The extension matters for GORP because orientation constraints are inherently directional, and directed templates $H$1 can encode admissible local or global orientation patterns.

Several technical objects are central in this framework. The paper uses a topological approach based on fundamental groupoids, together with mappings between walks in $H$2 and walks in $H$3, considered via reduction or equivalence classes. Associated to each recoloring sequence is a family of vertex walks recording the successive colors assumed by each vertex. This converts the original reconfiguration problem into a constrained walk problem on the template digraph.

The monochromatic neighborhood property is a key structural condition. When a vertex’s color is changed, all its neighbors must be colored identically; otherwise, forbidden substructures arise. The summary also refers to a monochromatic neighborhood or POP property. This condition is the mechanism that keeps local changes globally consistent, and it is one of the reasons the method extends to orientation-sensitive settings.

For GORP, the significance is not that a single-edge flip is literally the same as a single-vertex recoloring, but that feasible orientations can be represented by homomorphisms into a fixed digraph. Under such an encoding, orientation reconfiguration inherits the walk-based and groupoid-based machinery developed for $H$4-Recoloring. The cited work states explicitly that these topological and walk-based characterizations, especially the treatment of orientations and orientation-compatibility, can be directly applied to versions of GORP where feasible orientations correspond to homomorphisms into certain digraphs (Lévêque et al., 2022).

3. Structural tractability through forbidden cycles

A major contribution of the homomorphism approach is the identification of template classes $H$5 for which reconfiguration is solvable in polynomial time. The obstruction language is expressed through algebraic girth. For a directed cycle, its algebraic girth is the absolute value of the difference between the number of forward arcs and backward arcs in the cycle.

Two polynomial-time theorems are central.

Template class	Forbidden subgraphs	Consequence
Loopless digraph $H$6	No 4-cycle of algebraic girth 0	$H$7-Recoloring is solvable in polynomial time
Reflexive digraph $H$8	No triangle of algebraic girth 1 and no 4-cycle of algebraic girth 0	$H$9-Recoloring is solvable in polynomial time

The first theorem states that if $\phi: G \to H$0 is a loopless digraph with no 4-cycle of algebraic girth 0 as subgraph, then $\phi: G \to H$1-Recoloring is solvable in polynomial time. The second states that if $\phi: G \to H$2 is reflexive, has no triangle of algebraic girth 1, and no 4-cycle of algebraic girth 0, then $\phi: G \to H$3-Recoloring is solvable in polynomial time (Lévêque et al., 2022).

These results generalize previous polynomial cases for undirected graphs. The summary states that previous polynomial cases for undirected square-free graphs, that is, graphs with no 4-cycles, are extended to full directivity. It also states that for reflexive undirected graphs of girth at least 5, the algorithm applies as previously shown by Lee et al., 2021.

For GORP, the importance of these theorems is indirect but substantial. The stated connection is that polynomial-time results for classes of digraphs $\phi: G \to H$4 show that for some families of orientation constraints, GORP is tractable if the forbidden substructures are excluded from the template orientation digraph. This identifies algebraic-girth obstructions as a structural boundary for tractable orientation reconfiguration when orientations are represented through homomorphisms.

4. Algorithmic mechanisms for constructing reconfiguration sequences

The tractability results are algorithmic rather than purely existential. The procedure begins by reducing recoloring to a search for walks in $\phi: G \to H$5 corresponding to color-change sequences of vertices in $\phi: G \to H$6. The resulting candidate walks are then classified under the monochromatic neighborhood or POP property; the output of this classification is that there are either zero, one, or infinitely many possible walks.

Directedness introduces an additional compatibility constraint, called the zigzag condition. For digraphs, the successive colors in a vertex walk must be compatible with the arc directions in $\phi: G \to H$7. The summary gives the example that for a vertex $\phi: G \to H$8 with in-neighbors, the successive colors in its walk must “zigzag” in accordance with the arc structure. This is the point at which orientation sensitivity enters the reconfiguration analysis in an explicit way.

For reflexive digraphs, the Move Forward algorithm simulates feasible recoloring sequences by updating color states subject to the push/pull constraints from the POP property. For families of walks of the form $\phi: G \to H$9, it suffices to check only a polynomially bounded range of $G$0, by Lemma 27 and Lemma 28. The method also handles special cases through obstruction analysis, including the detection of frozen vertices via tight cycles and the presence or absence of symmetric walks (Lévêque et al., 2022).

The summary emphasizes that the polynomial-time guarantee is achieved by groupoid operations, breadth-first searches, and checking walks of bounded polynomial length. For GORP, this suggests an algorithmic template: encode feasible orientations into a digraph-homomorphism instance, classify allowable local moves by walk constraints, enforce orientation compatibility, and detect frozen regions through cycle obstructions. This is an inference from the described pipeline, but it follows closely from the role the paper assigns to orientation-compatibility and template-based tractability.

5. Local vertex constraints, SAT-E2 structure, and complexity dichotomies

A complementary perspective comes from graph orientation under local constraints. In "Planar Graph Orientation Frameworks, Applied to KPlumber and Polyomino Tiling" (Group et al., 3 Mar 2026), graph orientation is analyzed as a special kind of SAT problem in which each edge orientation is a variable and each variable appears in exactly two clauses, once positively and once negatively. This is the SAT-E2 connection.

The main non-planar dichotomy concerns sets $G$1 of symmetric vertex types. GO is in $G$2 if and only if all vertex types are bijunctive, affine, or have no “gap” of size $G$3 in their bitstring representation; otherwise, GO is NP-complete. With constants, the summary states that the only tractable cases are bijunctive, affine, or gap-free types. The planar case largely inherits the non-planar NP-hardness results, but two additional tractable classes appear: first, if all allowed vertex types are bottom-up closed or top-down closed; second, if $G$4 among allowed in-degrees is at least 5 and only trivial combinations are present, namely all-in, all-out, or only 1-in (Group et al., 3 Mar 2026).

The framework also introduces gadget classes, including duplicators, synchronizers, and alternators. Under singleton symmetric types and duplicators, a dichotomy is given: the problem is in $G$5 if all constraints are trivial or bijunctive, or the duplicator is an alternator, and NP-complete otherwise. The paper further gives a new polynomial-time algorithm for Planar GO with $G$6-in-$G$7 and $G$8-equalizer vertices when $G$9, resolving a 20-year-old open problem about KPlumber.

For GORP, the paper states that these results provide the foundation to classify the complexity of the decision version of GORP via the underlying orientation existence problem. It also states that for symmetric vertex-type classes where GO is in $H$0, it is plausible, though not always guaranteed, that the corresponding reconfiguration problems are also tractable, or at least that their solution spaces can be efficiently explored. Conversely, for NP-complete classes, the reconfiguration variant is also at least as hard as the decision variant, which provides a hardness basis (Group et al., 3 Mar 2026).

6. Scope, misconceptions, and research directions

One common misconception is to conflate orientation existence with orientation reconfiguration. The local-constraint framework yields sharp $H$1/NP-complete dichotomies for deciding whether a feasible orientation exists, but the summary is explicit that this does not automatically produce a full dichotomy for the corresponding reconfiguration problem. The appropriate formulation is weaker and more precise: the existence results provide a foundation, and tractability of reconfiguration is plausible in some polynomial cases, but not always guaranteed (Group et al., 3 Mar 2026).

A second misconception is that GORP is governed solely by local vertex constraints. The homomorphism-based results show that global template structure matters, especially forbidden cycles classified by algebraic girth. In the loopless and reflexive template classes singled out above, the absence of specific cycles is what permits the extension of Wrochna’s algorithm and the corresponding polynomial-time reconfiguration results. This places GORP in a broader structural theory where local move rules are constrained by global cycle obstructions (Lévêque et al., 2022).

A third misconception is that orientation reconfiguration and recoloring reconfiguration are merely analogous. The relationship stated in the literature is stronger: when $H$2 encodes orientation patterns or forbidden structures, GORP reduces to an $H$3-Recoloring problem, and the resulting theory generalizes previous results for reconfiguring acyclic orientations, colorings, and partial orders as special cases of digraph homomorphism reconfiguration (Lévêque et al., 2022). This does not make all orientation reconfiguration problems homomorphism problems in a trivial sense, but it identifies a substantial class that can be treated uniformly.

The present landscape therefore consists of two complementary structural programs. One is template-based and topological, centered on groupoids, walks, algebraic girth, the monochromatic neighborhood or POP property, and zigzag compatibility. The other is local and SAT-like, centered on symmetric vertex types, gap structure in bitstrings, closure properties, and gadget simulation. A plausible implication is that a mature theory of GORP will require combining these programs: template obstructions appear to govern global navigability of the reconfiguration space, while SAT-E2 and vertex-type dichotomies govern which feasible orientations exist in the first place.