Abstract: Self equivalences of classifying spaces of $p$-local compact groups are well understood by means of the algebraic structure that gives rise to them, but explicit descriptions are lacking. In this paper we use a construction of Robinson of an amalgam $G$, realizing a given fusion system, to produce a split epimorphism from the outer automorphism group of $G$ to the group of homotopy classes of self homotopy equivalences of the classifying space of the corresponding $p$-local compact group.
Introduction: The Problem of Symmetry in $p$-Local Compact Groups
In algebraic topology, $p$-local compact groups serve as sophisticated models for the $p$-completed homotopy theory of compact Lie groups. Defined by a triple—a Sylow-like subgroup ($S$), a fusion system ($\mathcal{F}$), and a linking system ($\mathcal{L}$)—these objects are often difficult to manipulate directly. A central challenge is understanding their self-equivalences: how these objects can be mapped to themselves while preserving their internal structure. While theoretical frameworks exist via classifying spaces ($BG$), explicit computational tools for these symmetries have been historically lacking. This research addresses this gap by relating topological symmetries to the automorphisms of a discrete group that "realizes" the underlying fusion system.
Methodology: The Robinson Amalgam and Split Epimorphism
The authors bridge the gap between abstract topology and discrete group theory using two primary techniques:
- Robinson Amalgams: Utilizing a construction by G. Robinson (refined by Libman and Seeliger), the authors build a discrete group $G$ as a colimit of a "tree of groups." This group $G$ captures the conjugacy relations (fusion) of the $p$-local compact group.
- Split Epimorphism Construction: The core contribution is the construction of a map $\omega: Out(G) \to Out(BG)$. The authors prove this map is a split epimorphism, ensuring that every symmetry of the classifying space $BG$ can be lifted to a symmetry of the discrete group $G$.
To handle the infinite nature of $p$-local compact groups (e.g., $p$-toral groups), the authors employ "bullet constructions" and unstable Adams operations to approximate these structures with sequences of finite $p$-local groups, making the analysis tractable.
Key Insights: Centers and Prime-Specific Behavior
- The Amalgam as a Proxy: The outer automorphism group of a complex, infinite-dimensional topological space ($BG$) can be effectively studied through the outer automorphisms of a discrete group ($G$).
- The Role of the Center: The "gap" between the symmetries of the amalgam and those of the $p$-local compact group—specifically the kernel of the map $\omega$—is intrinsically linked to the centers of the involved subgroups.
- Prime Disparity ($p=2$ vs. Odd Primes): The prime $p=2$ exhibits unique complexity. For odd primes, the relationship is straightforward. However, for $p=2$, a "higher limit" term ($\varprojlim^1$) appears in the short exact sequence, indicating that the topology at the prime 2 is more subtly "twisted."
Key Results
The paper culminates in several significant findings:
- Theorem A: Establishes the existence of the group $G$ and the split epimorphism $\omega: Out(G) \to Out(BG)$. It provides a precise description of the kernel of $\omega$ as a specific subgroup of automorphisms fixing the Sylow subgroup $S$.
- Proposition B: Identifies conditions (the Libman-Seeliger setup) under which $\omega$ becomes a full isomorphism, equating the algebraic and topological symmetries.
- Exotic Applications: The theory is applied to "exotic" $p$-local groups (e.g., exotic 3-local and 7-local groups), where these conditions hold, providing immediate computational utility.
Limitations and Future Directions
Despite its strengths, the Robinson amalgam $G$ can be difficult to describe explicitly for complex fusion systems. Additionally, the $\varprojlim^1$ obstruction at $p=2$ suggests remaining topological nuances that the current algebraic model does not fully simplify. Future research may explore alternative model groups to eliminate this term or generalize the framework to study all maps between different $p$-local compact groups, a major open problem in the field.
Summary updated