The complex of discrete Morse matchings of the n n -simplex: homotopy types and structural results

Nicholas A. Scoville · 2026

Read The complex of discrete Morse matchings of the n n -simplex: homotopy types and structural results by Nicholas A. Scoville (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-18172.txt.

Introduction

Scoville computes homotopy types for the complex of discrete Morse matchings on the 3-simplex and its boundary. These complexes collect acyclic matchings on Hasse diagrams, and their shapes were known in only a handful of cases. For general n, he reduces counting optimal matchings to the (n-2)-skeleton and shows the cone inclusion is null-homotopic. He also computes the f-vector for the 4-simplex, whose top entry of 380,125 counts its optimal discrete Morse matchings, and ends with two conjectures.

Catalog record

CREATORNicholas A. Scoville
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE67682
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.18172

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The complex of discrete Morse matchings of the n n -simplex: homotopy types and structural results Nicholas A. Scoville Abstract . The complex of discrete Morse matchings ℳ ⁡ ( K ) \mathcal{M}(K) , introduced by Chari and Joswig, is a simplicial complex whose simplices are the acyclic matchings on the Hasse diagram of K K . Its homotopy type is known in only a handful of cases. In this paper, we compute the homotopy types of ℳ ⁡ ( Δ 3 ) \mathcal{M}(\Delta^{3}) and ℳ ⁡ ( ∂ Δ 3 ) \mathcal{M}(\partial\Delta^{3}) , the corresponding pure complexes ℳ P ​ ( Δ 3 ) ≃ ℳ P ​ ( ∂ Δ 3 ) \mathcal{M}_{P}(\Delta^{3})\simeq\mathcal{M}_{P}(\partial\Delta^{3}) , and the generalized complex of discrete Morse matchings 𝒢 ​ ℳ ​ ( Δ 3 ) ≃ 𝒢 ​ ℳ ​ ( ∂ Δ 3 ) \mathcal{GM}(\Delta^{3})\simeq\mathcal{GM}(\partial\Delta^{3}) . For general n n we prove the identity f ⁡ ( n ) = ( n + 1 ) ⋅ | top-dimensional facets of ​ ℳ ​ ( Δ ( n − 2 ) n ) | f(n)=(n+1)\cdot|\text{top-dimensional facets of }\mathcal{M}(\Delta^{n}_{(n-2)})| , reducing the enumeration of optimal matchings on Δ n \Delta^{n} to an enumeration on its ( n − 2 ) (n-2) -skeleton, and we show that the inclusion ℳ ⁡ ( K ) ↪ ℳ ⁡ ( C ​ K ) \mathcal{M}(K)\hookrightarrow\mathcal{M}(CK) is null-homotopic for any cone. We also compute the f f -vector of ℳ ⁡ ( Δ 4 ) \mathcal{M}(\Delta^{4}) , whose top entry f ⁡ ( 4 ) = 380,125 f(4)=380{,}125 is the number of optimal discrete Morse matchings on Δ 4 \Delta^{4} . We conclude with two conjectures extending the ℳ P \mathcal{M}_{P} and 𝒢 ​ ℳ \mathcal{GM} equivalences to all n n . 2020 Mathematics Subject Classification Primary 05E45, 57Q70; Secondary 55P10, 05C70 1. Introduction Let K be a finite simplicial complex. The complex of discrete Morse matchings 1 1 1 In the literature, this has been called both the Morse complex and the complex of discrete Morse functions. The former name causes confusion with the name of the chain complex induced by the critical simplices of a discrete Morse function on K , while in the latter case, we agree with J. Brüggemann who points out that this construction refers “to a complex whose simplices correspond to [discrete] Morse matchings rather than [discrete] Morse functions” [ 4 , p. 26] . \mathcal{M}(K) was introduced by Chari and Joswig [ 7 ] . We may view \mathcal{M}(K) as a parameter space of gradient vector fields on K , or equivalently of discrete Morse matchings or acyclic matchings in the sense of [ 11 , 12 , 7 ] . While not every matching yields a substantive reduction, each simplex of \mathcal{M}(K) encodes a consistent set of cancellations and hence a choice of simplifying K up to simple homotopy. The topology of \mathcal{M}(K) reflects how flexible these choices are. For example, i -connectivity measures the extent to which families of matchings can be deformed through matchings, as studied in [ 18 ] , in analogy with Cerf-theoretic phenomena for spaces of smooth Morse functions [ 6 , 4 ] . Simplices of maximum dimension correspond to discrete Morse matchings that minimize the total number of critical simplices, and we call such matchings optimal . In favorable cases, one can achieve an optimal matching with the fewest critical simplices allowed by the discrete Morse inequalities [ 11 , Corollary 3.7] . Optimal and near-optimal matchings have important computational applications since discrete Morse reductions can substantially reduce complexes prior to homology or persistent homology computations in applied topology and topological data analysis [ 2 , 17 ] . Despite the central role of discrete Morse theory in combinatorial topology, the homotopy type of \mathcal{M}(K) is known only in a small number of cases. Kozlov computed the homotopy type of \mathcal{M}(K) for paths, cycles, and the complete graphs in [ 14 ] ; the homotopy type has been determined for several other families of graphs and special cases in [ 8 , 9 ] . The automorphism group of \mathcal{M}(K) has been computed in terms of the automorphism group of K in [ 16 ] , and Capitelli and Minian have shown that \mathcal{M}(K) encodes the isomorphism type of K in [ 5 ] . A natural next test case is the n -simplex \Delta^{n} and its boundary \partial\Delta^{n} . The homotopy type of \mathcal{M}(\Delta^{2}) is known from [ 7 ] , and even here very little is determined for n\geq 3 . The principal aim of this paper is to determine the homotopy types of \mathcal{M}(\Delta^{3}) , \mathcal{M}(\partial\Delta^{3}) , and their associated pure and generalized complexes. Along the way we develop structural tools that apply for all n and present computational data for n=4 . Our main results are for n=3 . Combining direct homology computations with the framework of Bravo and Camarena in [ 3 ] , we show in Section 5 that \mathcal{M}(\Delta^{3})\simeq\bigvee^{99}S^{4},\qquad\mathcal{M}(\partial\Delta^{3})\simeq\bigvee^{21}S^{3}\vee\bigvee^{24}S^{4}, and that the pure complexes in both cases share a single homotopy type in \mathcal{M}_{P}(\Delta^{3})\simeq\mathcal{M}_{P}(\partial\Delta^{3})\simeq\bigvee^{81}S^{3} . These appear as Theorems 20 , 21 , and 22 , respectively. In addition, we develop structural results that apply for all n . First, we construct an explicit bijection between the top-dimensional facets of \mathcal{M}(\Delta^{n}) and those of \mathcal{M}(\partial\Delta^{n}) in Proposition 9 . This implies that the two complexes admit the same number of optimal matchings. Second, we record a “layer count” for optimal matchings in Lemma 11 and use it to reduce their enumeration to an enumeration on the (n-2) -skeleton. This defines a surjection from optimal matchings on \Delta^{n} to optimal matchings on \Delta^{n}_{(n-2)} with uniform fiber size n+1 , yielding the identity f(n)=(n+1)\cdot\bigl|\text{top-dimensional facets of }\mathcal{M}(\Delta^{n}_{(n-2)})\bigr| where f(n) is the number of optimal matchings on \Delta^{n} (Theorem 14 ). Third, we prove in Theorem 17 that for any finite simplicial complex K , the inclusion \mathcal{M}(K)\hookrightarrow\mathcal{M}(CK) into the complex of discrete Morse matchings of the cone is null-homotopic. The proof involves a one-step contiguity using a single explicit Hasse edge of the cone, and when applied to the decomposition \Delta^{n}=v*\Delta^{n-1} it splits the long exact sequence of the pair (\mathcal{M}(\Delta^{n}),\mathcal{M}(\Delta^{n-1})) . For n=4 we present partial computational evidence. Exhaustive parallel enumeration gives the full f -vector of \mathcal{M}(\Delta^{4}) ; its top entry f(4)=380{,}125 agrees with the prediction of our restriction theorem via the identity f(4)=5\cdot 76{,}025 . The reduced Euler characteristic \widetilde{\chi}(\mathcal{M}(\Delta^{4}))=212{,}456 is positive, which suffices to rule out any homotopy equivalence of \mathcal{M}(\Delta^{4}) with a wedge of spheres in a single odd dimension. The full integral homology of \mathcal{M}(\Delta^{4}) remains out of reach, as the largest boundary matrix has dimensions on the order of 10^{9} , beyond what our present computational setup can handle. We close the paper with a comparison to the generalized complex of discrete Morse matchings \mathcal{GM}(K) of [ 18 ] , which drops the acyclicity condition and is the simplicial complex of all matchings on the Hasse diagram of K . Unlike for \mathcal{M} , where \mathcal{M}(\Delta^{3})\not\simeq\mathcal{M}(\partial\Delta^{3}) , we show in Section 6 that \mathcal{GM}(\Delta^{3})\simeq\mathcal{GM}(\partial\Delta^{3})\simeq\bigvee^{39}S^{4} , and we conjecture that \mathcal{GM}(\Delta^{n})\simeq\mathcal{GM}(\partial\Delta^{n}) for all n\geq 2 in Conjecture 28 . Together with the

Read the complete text