Homology groups of cubical sets with connections (Q2044555)

Cubical sets have some properties that do not correspond to our intuition. To correct this deficiency, \textit{R. Brown} and \textit{C. B. Spencer} [Cah. Topol. Gém. Différ. 17, 343--362 (1976; Zbl 0344.18004)] added extra degeneracy axioms, which led to the definition of \textit{cubical sets with connections} in [\textit{R. Brown} and \textit{P. J. Higgins}, J. Pure Appl. Algebra 21, 233--260 (1981; Zbl 0468.55007)]. The success of this approach includes: a proof of the fibrantness theorem, that a constrained cubical fibrant group with connections is Kan fibrant [\textit{A.P. Tonks}, J. Pure Appl. Algebra 81, No. 1, 83--87 (1992; Zbl 0757.55015)], a proof that the category of cubical Abelian groups with connections is equivalent to the category of chain complexes [\textit{R. Brown} and \textit{P. J. Higgins}, Homology Homotopy Appl. 5, No. 1, 49--52 (2003; Zbl 1025.18006)]. Also, in [\textit{G. Maltsiniotis}, ibid. 11, No. 2, 309--326 (2009; Zbl 1213.18009)], it was shown that cubical sets with connections form a strict test category, and it was pointed out that, in particular, this implies that cubical sets with connections are models for homotopy types in a very precise sense, in a way that is compatible with the Cartesian product. The article is devoted to the following question. Consider the singular cubical homology of a topological space, the same homology as calculated, for example, by \textit{J. P. Serre} [Ann. Math. (2) 54, 425--505 (1951; Zbl 0045.26003)]. Will this differ from the homology groups of a singular cubical set with connections? The authors claim that these groups are isomorphic. The connection chain subcomplex is constructed in Section 3. Let \(K\) be a cubical set with connections and let \({\mathcal C}(K)\) be its nondegenerate chain complex in the category of \(R\)-modules. Connections are denoted by \(\Gamma^{\beta}_i\). For \(n\geq 0\), let Con\(_n(K)\) be the \(R\)-submodule of \({\mathcal C}_n(K)\) which is generated by the cosets of \(\Gamma^{\beta}_i(\tau)\) where \(\tau\in K_{n-1}\) (Definition 5). It is proved that the homology of the complex Con\((K)\) is trivial for all \(n>0\) (Theorem 10). Corollary 12. The short exact sequence of chain complexes \[ 0 \to \mathrm{Con}_n(K) \hookrightarrow {\mathcal C}_{n+1}(K) \twoheadrightarrow{\mathcal C}_{n+1}(K)/\mathrm{Con}_n(K)\to 0 \] induces a long exact sequence of homology groups, and since \({\mathcal H}_n(\mathrm{Con}(K))\) is trivial, we have \({\mathcal H}_n({\mathcal C}(K))\cong {\mathcal H}_n({\mathcal C}(K)/\mathrm{Con}(K))\). The article also contains an appendix in which it is proved that homology modules of geometric realizations of a cubical set with connections are isomorphic to the homology modules of a geometric realization of a cubical set contained in it.