Polydiagonal Subspace in ℝⁿ

• Polydiagonal subspace is a linear subspace in ℝⁿ defined by enforcing identical or negated coordinate constraints, capturing synchrony and anti-synchrony phenomena.
• It offers a systematic taxonomy distinguishing synchrony subspaces, minimally tagged, fully tagged, and evenly tagged anti-synchrony subspaces, critical for analyzing network dynamics.
• Computational methods using constraint programming and coloring vectors enable efficient enumeration of these subspaces, with applications to graphs like cycles, Petersen, and Buckyball.

A polydiagonal subspace in $\mathbb{R}^n$ is a linear subspace specified by enforcing that certain subsets of coordinates are either identical (synchrony) or negatives of each other (anti-synchrony). These structures generalize the synchrony patterns corresponding to equitable partitions in graph theory, providing a rigorous algebraic framework to describe synchrony and anti-synchrony phenomena in coupled cell networks and weighted networks. The classification, enumeration, and invariance properties of polydiagonal subspaces are central to understanding network dynamics under linear and certain nonlinear couplings, and recent advances allow their systematic computation via constraint programming methods (Neuberger et al., 2024, Nijholt et al., 2022).

1. Formal Definition and Structure

Let $C = \{1, \ldots, n\}$ index the components (cells, nodes) of $\mathbb{R}^n$. A polydiagonal subspace is defined by a tagged partition $\mathcal{P}$ of $C$; this consists of a set partition $\mathcal{P} = \{P_1, \ldots, P_d\}$ together with a partial involution $*: \mathcal{P} \to \mathcal{P}$ where $P^* = Q \Rightarrow Q^* = P$, and at most one part $P_0$ satisfies $P_0^* = P_0$. The subspace is

$C = \{1, \ldots, n\}$0

If $C = \{1, \ldots, n\}$1 is trivial (no anti-pairs), $C = \{1, \ldots, n\}$2 is a synchrony subspace; otherwise, anti-synchrony relations are present. The polydiagonal formalism cleanly captures all possible synchrony/anti-synchrony structures in networks (Neuberger et al., 2024, Nijholt et al., 2022).

2. Classification: Synchrony and Anti-Synchrony Subspaces

Polydiagonal subspaces decompose into the following categories:

• Synchrony subspaces: No anti-pairs ($C = \{1, \ldots, n\}$3), i.e., only $C = \{1, \ldots, n\}$4 constraints. The corresponding class is combinatorially enumerated by the Bell numbers.
• Anti-synchrony subspaces: At least one anti-pair. These further subdivide:
  • Minimally tagged: Exactly one self-involuting block ($C = \{1, \ldots, n\}$5).
  • Fully tagged: Every block is either paired with another (anti-pair) or fixed.
  • Evenly tagged: Each block and its anti-partner have equal cardinality; equivalently, all elements $C = \{1, \ldots, n\}$6 satisfy $C = \{1, \ldots, n\}$7, meaning the all-ones vector is orthogonal to $C = \{1, \ldots, n\}$8.
  • This taxonomy is significant in dynamical systems where certain symmetry or sign-reversal constraints describe admissible invariant manifolds or flow-invariant subspaces (Nijholt et al., 2022).

3. Invariance under Linear Operators

A key structural property is the invariance of a polydiagonal subspace $C = \{1, \ldots, n\}$9 under a matrix $\mathbb{R}^n$0; that is, $\mathbb{R}^n$1. The invariance criterion reduces to verifying that the image of each basis vector (see §4) of $\mathbb{R}^n$2 is again a linear combination respecting the synchrony and anti-synchrony constraints.

Specifically, $\mathbb{R}^n$3 is $\mathbb{R}^n$4-invariant if for any $\mathbb{R}^n$5 and all $\mathbb{R}^n$6:

• If $\mathbb{R}^n$7, then $\mathbb{R}^n$8.
• If $\mathbb{R}^n$9, then $\mathcal{P}$0.
• If $\mathcal{P}$1, then $\mathcal{P}$2.

For coupled cell networks with Laplacians or adjacency matrices, this linear framework precisely characterizes flow-invariant synchrony and anti-synchrony patterns (Neuberger et al., 2024, Nijholt et al., 2022). When the network matrix has constant column sums and a simple leading eigenvector $\mathcal{P}$3 with all $\mathcal{P}$4, every $\mathcal{P}$5-invariant polydiagonal subspace is either a synchrony subspace containing $\mathcal{P}$6 or an evenly tagged anti-synchrony subspace not containing $\mathcal{P}$7. For Laplacians of weight-balanced digraphs, the only invariant anti-synchrony subspaces are evenly tagged ones (Nijholt et al., 2022).

4. Coloring Vectors and Computational Encoding

Polydiagonal subspaces correspond bijectively to integer coloring vectors $\mathcal{P}$8 under a canonical encoding. The construction proceeds as follows:

• The first nonzero $\mathcal{P}$9 is $C$0.
• If $C$1 has appeared, a subsequent $C$2 may be $C$3 (anti-class) or $C$4 (new class).
• For each $C$5 in the set $C$6, define the sign pattern

$C$7

The collection $C$8 forms a basis for $C$9. The $\mathcal{P} = \{P_1, \ldots, P_d\}$0 matrix $\mathcal{P} = \{P_1, \ldots, P_d\}$1 with these vectors as columns gives $\mathcal{P} = \{P_1, \ldots, P_d\}$2. The coloring vector formulation provides a compact, computationally tractable representation of polydiagonal patterns and thoroughly encodes both synchrony and anti-synchrony constraints (Neuberger et al., 2024).

5. Constraint Programming Formulation and Complexity

Determining all polydiagonal subspaces invariant under a given $\mathcal{P} = \{P_1, \ldots, P_d\}$3 is equivalent to solving a Constraint Satisfaction Problem (CSP) over integer variables $\mathcal{P} = \{P_1, \ldots, P_d\}$4. Constraints comprise:

• Coloring vector validity (prefix and sign rules, e.g., $\mathcal{P} = \{P_1, \ldots, P_d\}$5, with $\mathcal{P} = \{P_1, \ldots, P_d\}$6).
• Invariance constraints: For each $\mathcal{P} = \{P_1, \ldots, P_d\}$7 and pairs $\mathcal{P} = \{P_1, \ldots, P_d\}$8,
  • $\mathcal{P} = \{P_1, \ldots, P_d\}$9,
  • $*: \mathcal{P} \to \mathcal{P}$0,
  • $*: \mathcal{P} \to \mathcal{P}$1,
  • where $*: \mathcal{P} \to \mathcal{P}$2.
  • All variable domains are finite, and all constraints are linear equalities, inequalities, or guarded equalities. The CSP can enumerate all valid coloring vectors (equivalently, all tagged partitions and polydiagonal subspaces) and identify those which are $*: \mathcal{P} \to \mathcal{P}$3-invariant (Neuberger et al., 2024).

Empirical evaluation demonstrates that:

• Enumerating all coloring vectors in $*: \mathcal{P} \to \mathcal{P}$4 grows as the Dowling numbers $*: \mathcal{P} \to \mathcal{P}$5 (OEIS A007405).
• Modern constraint solvers (e.g., DOcplex, ORtools CP-SAT) can compute all invariant subspaces for $*: \mathcal{P} \to \mathcal{P}$6 in minutes, compared to hours or infeasibility for earlier graph algorithms.
• The approach enables for the first time the practical computation of all invariant anti-synchrony subspaces for medium-sized networks; for instance, finding all 1,244 invariant polydiagonal subspaces in the Buckyball graph (60 vertices) is feasible in under 7 hours (Neuberger et al., 2024).

6. Enumeration and Combinatorial Results

The total number $*: \mathcal{P} \to \mathcal{P}$7 of polydiagonal subspaces in $*: \mathcal{P} \to \mathcal{P}$8 is given by the Dowling numbers and has exponential generating function

$*: \mathcal{P} \to \mathcal{P}$9

The number of synchrony subspaces is $P^* = Q \Rightarrow Q^* = P$0 (the Bell numbers). Anti-synchrony subspaces, including the evenly tagged and minimally/fully tagged types, are counted via generating functions: $P^* = Q \Rightarrow Q^* = P$1 where $P^* = Q \Rightarrow Q^* = P$2 is the modified Bessel function of order $P^* = Q \Rightarrow Q^* = P$3 (Nijholt et al., 2022). This combinatorial framework grounds the enumeration of polydiagonal patterns and highlights the rapid growth of possible subspaces even for modest $P^* = Q \Rightarrow Q^* = P$4.

7. Illustrative Examples and Applications

Several case studies illustrate the theory and computational methodology:

• Toy $P^* = Q \Rightarrow Q^* = P$5 digraph: For $P^* = Q \Rightarrow Q^* = P$6, there are six coloring vectors and hence six polydiagonal subspaces, corresponding to the lattice structure displayed in (Neuberger et al., 2024).
• Cycle graphs ($P^* = Q \Rightarrow Q^* = P$7) and their Laplacians: The CSP approach efficiently computes all invariant synchrony and polydiagonal subspaces up to $P^* = Q \Rightarrow Q^* = P$8, beyond the reach of previous 'split-and-cir' algorithms.
• Petersen graph: All 240 invariant polydiagonal subspaces are revealed in seconds and partitioned into 22 orbits; 147 of these are anti-synchrony subspaces.
• Buckyball (C$P^* = Q \Rightarrow Q^* = P$9) graph: All 1,244 invariant polydiagonal subspaces (340 synchrony) are found; iterative quotienting enables discovering nested invariants within synchrony subspaces. This methodology is applicable to network symmetry, stability of coupling patterns, and the study of emergent collective behaviors in large networks (Neuberger et al., 2024).

---

Polydiagonal subspaces constitute the complete algebraic characterization of exact synchrony and anti-synchrony in finite-dimensional vector spaces, endowed with a rich combinatorial and linear algebraic structure. Their encoding via coloring vectors and enumeration through CSP methods provides a concrete, computationally viable approach for analyzing high-dimensional network dynamics, graph symmetry, and invariant patterns (Neuberger et al., 2024, Nijholt et al., 2022).