ON FACTORIZATIONS OF THE SUBGROUPS OF SELF-HOMOTOPY EQUIVALENCES

Yi-Yun Shi, Hao ZhaoJournal of the Korean Mathematical Society 2008 · DOI

Abstract

For a pointed space $X$, the subgroups of self-homotopy equivalences $Aut_{\sharp}_N(X)$, $Aut_{\Omega}(X)$, $Aut_*(X)$ and $Aut_{\Sigma}(X)$ are considered. Here, $Aut_{\sharp}_N(X)$ is the group of all self-homotopy classes $f$ of $X$ such that $f_{\sharp}=id$ on homotopy groups $\pi_i(X)$ for $i \leq N$; $Aut_{\Omega}(X)$ consists of classes where the induced map on the loop space $\Omega f$ is the identity; $Aut_*(X)$ consists of classes inducing the identity on homology groups $H_i(X)$; and $Aut_{\Sigma}(X)$ consists of classes where the suspension $\Sigma g$ is the identity. This paper demonstrates that $Aut_{\Omega}(X_1 \times \dots \times X_n)$ admits two factorizations similar to those of $Aut_{\sharp}_N$, and that $Aut_{\Sigma}$ and $Aut_*$ on wedge sums $X_1 \vee \dots \vee X_n$ possess dual factorizations.

Introduction

In the field of algebraic topology, the group of self-homotopy equivalences of a space $X$, denoted $Aut(X)$, represents the fundamental symmetries of that space—specifically, the set of all maps from $X$ to itself that can be continuously deformed back to the identity. While $Aut(X)$ is a central object of study, its complete structure is often notoriously difficult to calculate.

To make progress, mathematicians focus on specific subgroups of $Aut(X)$ that preserve certain topological invariants. This paper, published in the Journal of the Korean Mathematical Society, explores the structural decomposition of these subgroups when the space $X$ is composed of smaller building blocks, such as Cartesian products ($X \times Y$) or wedge sums ($X \vee Y$). Authors Yi-Yun Shi and Hao Zhao show that these complex symmetry groups can be "factorized" into simpler components, much like a matrix can be decomposed into triangular parts.

Defining the Subgroups

The research focuses on four primary subgroups of self-homotopy equivalences:

  1. $Aut_{\sharp}_N(X)$: Maps that induce the identity on homotopy groups up to dimension $N$.
  2. $Aut_{\Omega}(X)$: Maps that induce the identity on the loop space of $X$.
  3. $Aut_*(X)$: Maps that induce the identity on the homology groups of $X$.
  4. $Aut_{\Sigma}(X)$: Maps that induce the identity on the suspension of $X$.

Previous research by Pavešić established that for product spaces, $Aut_{\sharp}_N(X)$ could be factorized. Shi and Zhao extend this logic to the loop, homology, and suspension subgroups, providing a more comprehensive map of how these symmetries behave under standard topological constructions.

The Engine of Factorization: Reducibility

The core theoretical requirement for these factorizations is the concept of reducibility. A self-map of a composite space (like $X \times Y$) is considered reducible if it can be decomposed into maps that interact predictably with the individual components.

The authors demonstrate that for the subgroups $Aut_{\Omega}$, $Aut_*$, and $Aut_{\Sigma}$, every self-equivalence is reducible. This is a significant finding because it implies that the "global" symmetries of a complex space are entirely determined by the "local" symmetries of its constituent parts and the maps between them. This property does not hold for the full group $Aut(X)$, which explains why these specific subgroups are more mathematically tractable.

Main Results

Factorization of Product Spaces

For a product space $X = X_1 \times \dots \times X_n$, the authors prove that $Aut_{\Omega}(X)$ can be uniquely factorized into a product of $n$ specific subgroups (Theorem 2.3). Furthermore, they establish an "LU-style" factorization (Theorem 2.4), drawing a direct analogy to linear algebra where a matrix is decomposed into Lower and Upper triangular matrices. This suggests that the symmetry group has a hierarchical structure where information flows between components in a restricted, manageable way.

Duality in Wedge Sums

One of the paper's most elegant contributions is the application of Eckmann-Hilton duality. In topology, the product ($X \times Y$) and the wedge sum ($X \vee Y$) are dual constructions, as are the loop functor ($\Omega$) and the suspension functor ($\Sigma$).

Shi and Zhao prove that for wedge sums of simply-connected CW-complexes, the homology-preserving group $Aut_*$ and the suspension-preserving group $Aut_{\Sigma}$ admit factorizations that are exactly dual to the product space results. This symmetry reinforces the deep structural links between different ways of measuring a space's shape.

Triviality and H-Spaces

The paper also examines H-spaces—spaces equipped with a continuous multiplication (like Lie groups). They prove that if $X$ is an H-space, then $Aut_{\Sigma}(X)$ is trivial (Corollary 4.2). This means that for these highly structured spaces, the only map that preserves the suspension is the identity map itself, providing a sharp limit on the possible symmetries of H-spaces.

Conclusion and Future Directions

While the results for product spaces are broadly applicable, the authors note that the wedge sum factorizations require the spaces to be simply-connected CW-complexes. This leaves open the question of how these groups behave in more complex, non-simply-connected environments.

The paper concludes with two provocative conjectures:

  1. The existence of a finite CW-complex where $Aut_{\Sigma}(X)$ is trivial, but $Aut_{\Omega}(X)$ is not, which would further distinguish the behavior of these two dual functors.
  2. A proposed bound for the "nilpotency" of $Aut_{\Sigma}(X)$ based on the Lusternik-Schnirelmann cocategory, a measure of a space's geometric complexity.

By providing these factorization theorems, Shi and Zhao have given topologists a systematic toolkit for breaking down the symmetries of complex spaces into their simplest algebraic atoms.

Summary updated

Abstract

For a pointed space X, the subgroups of self-homotopy equivalences <TEX>$Aut_{\sharp}_N(X)$</TEX>, <TEX>$Aut_{\Omega}(X)$</TEX>, <TEX>$Aut_*(X)$</TEX> and <TEX>$Aut_{\Sigma}(X)$</TEX> are considered, where <TEX>$Aut_{\sharp}_N(X)$</TEX> is the group of all self-homotopy classes f of X such that <TEX>$f_{\sharp}=id\;:\;{\pi_i}(X){\rightarrow}{\pi_i}(X)$</TEX> for all <TEX>$i{\leq}N{\leq}{\infty}$</TEX>, <TEX>$Aut_{\Omega}(X)$</TEX> is the group of all the above f such that <TEX>${\Omega}f=id;\;Aut_*(X)$</TEX> is the group of all self-homotopy classes g of X such that <TEX>$g_*=id\;:\;H_i(X){\rightarrow}H_i(X)$</TEX> for all <TEX>$i{\leq}{\infty}$</TEX>, <TEX>$Aut_{\Sigma}(X)$</TEX> is the group of all the above g such that <TEX>${\Sigma}g=id$</TEX>. We will prove that <TEX>$Aut_{\Omega}(X_1{\times}\cdots{\times}X_n)$</TEX> has two factorizations similar to those of <TEX>$Aut_{\sharp}_N(X_1{\times}\cdots{\times}\;X_n)$</TEX> in reference [10], and that <TEX>$Aut_{\Sigma}(X_1{\vee}\cdots{\vee}X_n)$</TEX>, <TEX>$Aut_*(X_1{\vee}\cdots{\vee}X_n)$</TEX> also have factorizations being dual to the former two cases respectively.
BibTeX
@article{shi2008factorizations,
  author = {Yi-Yun Shi and Hao Zhao},
  title = {{ON FACTORIZATIONS OF THE SUBGROUPS OF SELF-HOMOTOPY EQUIVALENCES}},
  journal = {Journal of the Korean Mathematical Society},
  year = {2008},
  doi = {10.4134/jkms.2008.45.4.1089},
}
Figure