Minimum-Weight Parity Factor (MWPF)

• MWPF is an optimization problem defined to select a minimum-weight set of edges whose incidence parity exactly matches a prescribed defect set.
• It bridges graph parity factors and weighted linear matroid parity through Pfaffian formulations and primal-dual augmentation algorithms.
• The approach generalizes to hypergraphs via the HyperBlossom framework, providing certified decoding methods for quantum error correction.

Searching arXiv for recent and foundational papers on Minimum-Weight Parity Factor and related matroid-parity formulations. Minimum-Weight Parity Factor (MWPF) is an optimization problem that asks for a minimum-weight subset of edges or hyperedges whose incidence parity realizes a prescribed defect set. In graph-theoretic form, it subsumes parity-constrained subgraph selection such as $T$-joins; in hypergraph form, it provides a decoding objective for quantum error correction; and in matroidal form, it can be expressed as a weighted linear matroid-parity problem on an appropriate representation matrix. The modern literature presents MWPF along two closely connected lines: a reduction of graph parity factors to weighted linear matroid parity via Pfaffian formulations and primal-dual augmentation (Iwata et al., 2019), and a hypergraph generalization in which MWPF is cast as a primal-dual covering problem over invalid subgraphs, leading to the HyperBlossom framework for decoding quantum LDPC codes (Wu et al., 7 Aug 2025).

1. Definition and core formulations

In the hypergraph formulation, $G=(V,E)$ is a (hyper)graph, $D\subseteq V$ is the syndrome or defect set, and $w:E\to\mathbb{R}_+$ assigns non-negative weights to hyperedges. A subset $\mathcal{E}\subseteq E$ is a parity factor for syndrome $D$ if the induced defect set

$$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$

satisfies $\mathcal D(\mathcal E)=D$. The MWPF problem is then

$$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$

Equivalently, with binary variables $x_e\in\{0,1\}$, one minimizes $G=(V,E)$0 subject to parity constraints at every vertex: odd parity on $G=(V,E)$1, even parity on $G=(V,E)$2 (Wu et al., 7 Aug 2025).

In graphs, MWPF coincides with finding a minimum-weight subgraph whose vertex degrees have prescribed parities. The exposition based on weighted linear matroid parity states that graph-parity factors ($G=(V,E)$3-joins) are a special case on the cographic matroid of an undirected graph $G=(V,E)$4. After fixing an orientation, duplicating each undirected edge into two oriented columns, and pairing those columns into lines $G=(V,E)$5, a parity base in the associated linear matroid corresponds exactly to a set $G=(V,E)$6 of edges so that each vertex has the prescribed parity of incident edges (Iwata et al., 2019).

The weighted linear matroid-parity problem itself is defined on a full-row-rank matrix $G=(V,E)$7, with the ground set given by the columns of $G=(V,E)$8, partitioned into disjoint lines $G=(V,E)$9, each line $D\subseteq V$0 carrying a real weight $D\subseteq V$1. A base $D\subseteq V$2 in the linear matroid $D\subseteq V$3 is called a parity base if it is a union of lines. The optimization problem is

$D\subseteq V$4

with optimal value denoted $D\subseteq V$5 (Iwata et al., 2019). In this sense, graph MWPF is embedded into a more general algebraic parity-selection problem.

2. Reduction from graph parity factors to weighted linear matroid parity

The reduction described for graphs begins with the oriented incidence matrix of an undirected graph $D\subseteq V$6, with one row deleted so that the matrix has full rank. Each undirected edge $D\subseteq V$7 is duplicated into two columns corresponding to the two orientations, producing a matrix $D\subseteq V$8 of dimension $D\subseteq V$9. The line system is then $w:E\to\mathbb{R}_+$0, and the line weight is $w:E\to\mathbb{R}_+$1 (Iwata et al., 2019).

Under this construction, a parity base in $w:E\to\mathbb{R}_+$2 corresponds exactly to choosing a set of edges whose incidence parity matches the prescribed vertex parity pattern. The exposition states this equivalence directly: a parity base in $w:E\to\mathbb{R}_+$3 corresponds exactly to choosing a set $w:E\to\mathbb{R}_+$4 of edges so that each vertex has the prescribed parity of incident edges, i.e. $w:E\to\mathbb{R}_+$5 is a $w:E\to\mathbb{R}_+$6-join or parity factor (Iwata et al., 2019). Consequently, MWPF in graphs reduces in one shot to a weighted linear-matroid-parity instance $w:E\to\mathbb{R}_+$7.

This reduction is important because it places parity-factor optimization inside the well-developed theory of linear matroid parity. The underlying weighted linear matroid-parity algorithm of Iwata and Kobayashi is described as a combinatorial, deterministic, polynomial-time algorithm that builds on a polynomial matrix formulation using Pfaffians and adopts a primal-dual approach based on the augmenting path algorithm of Gabow and Stallmann for the unweighted problem (Iwata et al., 2019). This establishes a direct methodological bridge from blossom-style graph algorithms to algebraic matroid optimization.

A plausible implication is that MWPF inherits both combinatorial and algebraic viewpoints: combinatorially through parity-constrained subgraph selection, and algebraically through bases, lines, and exchange structures in represented matroids.

3. Pfaffian polynomial formulation and min-max structure

A central algebraic formulation introduces an indeterminate $w:E\to\mathbb{R}_+$8 and indeterminates $w:E\to\mathbb{R}_+$9 for each line $\mathcal{E}\subseteq E$0. One forms the skew-symmetric polynomial matrix

$\mathcal{E}\subseteq E$1

where

$\mathcal{E}\subseteq E$2

Its Pfaffian expands over perfect matchings in the support graph, and Murata’s combinatorial relaxation lemma yields the degree identity

$\mathcal{E}\subseteq E$3

Equivalently, the highest exponent of $\mathcal{E}\subseteq E$4 in the Pfaffian recovers the optimum weight of the weighted linear matroid-parity problem (Iwata et al., 2019).

The exposition makes the combinatorial meaning explicit: $\mathcal{E}\subseteq E$5 carries the weight of the complement of a parity base, so reading off the highest exponent recovers $\mathcal{E}\subseteq E$6 (Iwata et al., 2019). For graph MWPF, this Pfaffian representation provides an indirect encoding of parity-feasible edge sets through a perfect-matching expansion in an auxiliary skew-symmetric matrix.

The same source also presents a min-max characterization. Although no explicit LP of polynomial size is written in the paper, the Pfaffian-matching can be viewed as a max-weight perfect matching in an auxiliary graph $\mathcal{E}\subseteq E$7, whose edge weights are the $\mathcal{E}\subseteq E$8-degrees of corresponding entries after row and column operations. The standard matching-LP dual induces dual variables $\mathcal{E}\subseteq E$9 and $D$0, with constraints of the form $D$1 on each line and slack inequalities on non-base exchange pairs (Iwata et al., 2019).

The key inequality for an exchange pair $D$2 with $D$3 is

$D$4

Edges satisfying equality are called tight, and a matching is tight if all its edges are tight and no positive-$D$5 blossom is crossed more than once (Iwata et al., 2019). Standard LP duality then yields the min-max relation between minimum parity-base weight and the maximum feasible dual objective, with complementary slackness characterizing optimal primal-dual pairs.

This structure places MWPF, at least in its graphic and represented-matroid incarnations, within the general primal-dual tradition of matching theory. The formulation is not merely existential: it supports an augmenting-path algorithm with blossom management, tight-edge search, and certified optimality.

4. Primal-dual augmenting-path algorithm in the linear matroid setting

The weighted linear matroid-parity algorithm maintains a current base $D$6, dual variables $D$7 and $D$8, and a laminar family of blossoms $D$9, each with a Boolean variable indicating whether it is positive, i.e. whether $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$0 (Iwata et al., 2019). Initialization arbitrarily splits each line weight $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$1 into $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$2, then runs a greedy matroid-base algorithm on $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$3 to find a minimum-$$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$4 base $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$5, sets $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$6, and starts with $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$7.

The main loop proceeds until $$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$8 is a parity base. First, one builds the equality graph

$$\mathcal D(\mathcal E)=\{\,v\in V\mid |\{e\in\mathcal E:e\ni v\}|\equiv 1 \pmod 2\}$$9

and writes $\mathcal D(\mathcal E)=D$0 (Iwata et al., 2019). Second, one performs a blossom-style alternating BFS in $\mathcal D(\mathcal E)=D$1, labeling singletons in source lines as roots, growing alternating search trees, and creating new blossoms whenever an outer-outer collision occurs. If a path $\mathcal D(\mathcal E)=D$2 is found from a source vertex in one source line to a source vertex in a different source line, augmentation is triggered.

If no augmenting path is found, the algorithm performs a dual update. The update computes

$\mathcal D(\mathcal E)=D$3

then shifts $\mathcal D(\mathcal E)=D$4 on $\mathcal D(\mathcal E)=D$5 and $\mathcal D(\mathcal E)=D$6 according to base membership and adjusts $\mathcal D(\mathcal E)=D$7 by $\mathcal D(\mathcal E)=D$8 or $\mathcal D(\mathcal E)=D$9 on designated blossoms. Any blossom with $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$0 is immediately expanded (Iwata et al., 2019).

When an augmenting path exists, the algorithm writes the path as a tight perfect-matching-alternating sequence and augments via a symmetric-difference pivot. For each positive blossom crossed by the path, new bud and tip vertices are introduced to maintain the blossom-tree invariants, $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$1 is updated by a 4-way pivot, and $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$2 is adjusted accordingly. Then $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$3 is pivoted around all vertices of the path at once, giving the new base $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$4. Blossoms that become trivial are collapsed, and routing in each remaining blossom is recomputed recursively (Iwata et al., 2019).

The exposition emphasizes that throughout the procedure one maintains structural conditions $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$5 on $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$6 and dual-feasibility conditions $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$7, and that optimality of the final parity base follows from the Pfaffian min-max argument, specifically Theorem 4.1 in the paper (Iwata et al., 2019). In the specialized graph setting, these operations correspond to the familiar grow, shrink, dual-update, and alternating-tree manipulations of blossom-based parity algorithms.

5. Hypergraph MWPF and the HyperBlossom framework

The 2025 formulation generalizes MWPF from graphs to hypergraphs and is motivated by quantum error correction decoding (Wu et al., 7 Aug 2025). The central obstacle is that, for general hypergraphs, the natural LP relaxation need not be tight: the source states explicitly that for general hypergraphs $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$8, so solving the LP to optimality and rounding $$\min_{\mathcal E\subseteq E,\;\mathcal D(\mathcal E)=D} W(\mathcal E)=\sum_{e\in \mathcal E} w_e.$$9 may not give an exact MWPF (Wu et al., 7 Aug 2025).

To address this, the problem is reformulated as a covering-type ILP over invalid subgraphs. A subgraph $x_e\in\{0,1\}$0, with $x_e\in\{0,1\}$1 and $x_e\in\{0,1\}$2, is invalid if no choice $x_e\in\{0,1\}$3 realizes the local syndrome $x_e\in\{0,1\}$4. Denoting by $x_e\in\{0,1\}$5 the collection of all invalid subgraphs, one writes the cover-based ILP

$x_e\in\{0,1\}$6

with $x_e\in\{0,1\}$7, where $x_e\in\{0,1\}$8 is the “hair” of $x_e\in\{0,1\}$9, namely the edges incident to $G=(V,E)$00 but not in $G=(V,E)$01 (Wu et al., 7 Aug 2025). Relaxing integrality to $G=(V,E)$02 gives an LP with dual

$G=(V,E)$03

Complementary slackness takes the form

$G=(V,E)$04

HyperBlossom is presented as a primal-dual method that drives the gap to zero or certifies proximity (Wu et al., 7 Aug 2025). This is the hypergraph analogue of blossom-style exactness certificates in graph matching, but with dual variables defined on invalid subgraphs rather than only on odd-cardinality vertex sets.

The algorithm alternates two phases. The dual phase maintains a feasible dual vector $G=(V,E)$05 and identifies tight edges

$G=(V,E)$06

The primal phase operates cluster-by-cluster, where clusters are connected components of $G=(V,E)$07 acting on $G=(V,E)$08, with $G=(V,E)$09 the set of hyperblossoms (Wu et al., 7 Aug 2025). For a cluster $G=(V,E)$10, the primal phase first tests local optimality by solving MWPF on $G=(V,E)$11; if the local optimum $G=(V,E)$12 equals $G=(V,E)$13, the cluster is locally optimal. Otherwise it calls $G=(V,E)$14 to extract relaxers $G=(V,E)$15, removes the relaxed edges, and if the residual subgraph is still invalid, chooses a trivial dual direction $G=(V,E)$16 supported on an invalid subgraph (Wu et al., 7 Aug 2025).

A relaxer is defined as a dual direction that remains feasible and strictly decreases the tightness of at least one tight edge. The paper describes a general framework for batch relaxing by composing multiple relaxers into one $G=(V,E)$17 (Theorem 2) (Wu et al., 7 Aug 2025). For general hypergraphs, the SingleHair subroutine constructs a hair matrix $G=(V,E)$18 for each current hyperblossom $G=(V,E)$19, looks for an odd row, and from that row derives an invalid super-subgraph $G=(V,E)$20 with a valid relaxer $G=(V,E)$21, $G=(V,E)$22. If no odd row exists for any $G=(V,E)$23, the dual is in a “single-hair” state and no further relaxer is found (Wu et al., 7 Aug 2025).

The dual phase then adds any new blossom with $G=(V,E)$24 to the cluster’s dual history, re-solves the partial LP on variables in that history to optimality, recomputes $G=(V,E)$25 and $G=(V,E)$26, and re-merges clusters (Wu et al., 7 Aug 2025). The top-level pseudocode repeatedly alternates these steps until all clusters are locally optimal, at which point it returns the global parity factor $G=(V,E)$27 together with the dual vector $G=(V,E)$28.

6. Complexity, specialization, and decoding applications

For weighted linear matroid parity, the running-time analysis in the exposition yields the following bounds. Let $G=(V,E)$29 and $G=(V,E)$30. Each blossom-oriented search, including creation of new blossoms and maintenance of $G=(V,E)$31, takes $G=(V,E)$32 field operations; each dual update also takes $G=(V,E)$33; there are at most $G=(V,E)$34 dual updates between successive augmentations by Lemma 7.2; each augmentation takes $G=(V,E)$35; and there are at most $G=(V,E)$36 augmentations. The resulting total is $G=(V,E)$37. Over a fixed finite field, this gives a strongly-polynomial $G=(V,E)$38-time algorithm (Iwata et al., 2019).

In the graphic specialization, $G=(V,E)$39 and $G=(V,E)$40. The exposition states that one need not explicitly build an $G=(V,E)$41 skew-symmetric matrix: cavity-pivot operations on $G=(V,E)$42 correspond exactly to the usual grow, shrink, and dual-update steps of $G=(V,E)$43-join via cographic blossoms, while blossom creation and expansion correspond to shrinking odd cuts in the cographic matroid (Iwata et al., 2019). The resulting bound is $G=(V,E)$44 time for obtaining a minimum-weight subgraph with the prescribed vertex-degree parities (Iwata et al., 2019).

For HyperBlossom on general hypergraphs, worst-case complexity is described in terms of the number $G=(V,E)$45 of invalid subgraphs and the cost $G=(V,E)$46 of one relaxer-finder call. One relaxation round calls at most $G=(V,E)$47 relaxers, giving $G=(V,E)$48; there are at most $G=(V,E)$49 rounds until no more new blossoms; each dual phase solves an LP in $G=(V,E)$50 variables; and summing over all rounds gives a worst-case $G=(V,E)$51 by an interior-point bound. The total is therefore bounded by

$G=(V,E)$52

(Wu et al., 7 Aug 2025).

The same work emphasizes a certified proximity property: at any time, primal and dual feasible solutions satisfy

$G=(V,E)$53

and the true optimum lies in this gap; when the gap is zero, exact optimality is certified (Wu et al., 7 Aug 2025). This certificate is especially important in decoding settings, where heuristic decoders often lack verifiable suboptimality bounds.

HyperBlossom is positioned as a unified framework for quantum error correction decoding. The paper states that it formulates Most-Likely-Error decoding as an MWPF problem and generalizes the blossom algorithm to hypergraphs via a similar primal-dual linear programming model with certifiable proximity bounds (Wu et al., 7 Aug 2025). It further states that HyperBlossom unifies existing graph-based decoders such as (Hypergraph) Union-Find decoders and the Minimum-Weight Perfect Matching decoder (Wu et al., 7 Aug 2025). The implementation, Hyperion, is described as Rust software with both floating and rational LP and Python bindings, and the reported empirical results include a 4.8x lower logical error rate than MWPM on the distance-11 surface code, a 1.6x lower logical error rate than a fine-tuned BPOSD decoder on the $G=(V,E)$54 bivariate bicycle code under code-capacity noise, and almost-linear average runtime scaling on the surface code and color code up to code distances 99 and 31 for code-capacity and circuit-level noise, respectively (Wu et al., 7 Aug 2025).

A common misconception is that MWPF is merely a renaming of MWPM. The hypergraph formulation makes clear that classical MWPM on simple graphs is only the special case of “degree-2 only hyperedges” (Wu et al., 7 Aug 2025). Another common misconception is that blossom methods intrinsically require graph matching structure. The recent literature suggests instead that blossom-like primal-dual logic can be lifted to broader parity-constrained settings, although exactness in the hypergraph case depends on additional structure or on gap certification rather than blanket LP integrality (Wu et al., 7 Aug 2025).

7. Conceptual significance and relation between the two lines of work

The two arXiv sources describe MWPF from different but compatible perspectives. The 2019 weighted linear matroid-parity work provides the algebraic and combinatorial machinery for exact minimum-weight parity selection in represented matroids, together with a Pfaffian min-max theory and an augmenting-path algorithm (Iwata et al., 2019). The 2025 HyperBlossom work extends the parity-factor viewpoint to hypergraphs and uses a covering-dual formulation over invalid subgraphs to obtain a certifying primal-dual algorithm for decoding (Wu et al., 7 Aug 2025).

Their common structure is the use of tight objects, dual feasibility, and blossom-like contractions or generalized relaxations. In the matroid setting, tightness is defined by equality in exchange inequalities involving $G=(V,E)$55 and $G=(V,E)$56, and blossoms are laminar structures governing augmenting-path search (Iwata et al., 2019). In the hypergraph setting, tightness is defined by saturated dual constraints $G=(V,E)$57, hyperblossoms are invalid subgraphs with positive dual weight, and relaxers play the role of generalized dual directions that expose new structure (Wu et al., 7 Aug 2025).

This suggests a broad interpretation of MWPF as a parity-constrained optimization paradigm rather than a single narrow graph problem. In graphs, it recovers $G=(V,E)$58-joins through cographic matroids; in represented matroids, it becomes weighted linear matroid parity; in hypergraphs, it supports a certifying decoder architecture for qLDPC codes (Iwata et al., 2019, Wu et al., 7 Aug 2025). The continuity across these settings lies in a recurring min-max pattern: a primal object encoding parity feasibility, a dual object measuring obstruction or invalidity, and an algorithmic mechanism that alternates structural search with dual adjustment until optimality or certified proximity is established.