AI 中文总结
该研究围绕单位圆上的$\u27e8mathcal{I}\u27e9$-特征化子群，结合理想的结构与拓扑性质，刻画其描述复杂度与可Polish化性，得到解析P-理想下的可Polish化等价刻画、复杂度三分性等结果，并给出具体序列对应的子群描述与开放问题。
AI 中文摘要
给定ω上的理想$\u27e8mathcal{I}\u27e9$，若存在整数序列$(a_n:n\in\omega)$，使得单位圆$\u27e8mathbb{T}\u27e9$的子群$H$满足$$ H= \left\{ x\in\mathbb{T}: \mathcal{I}\text{-}\lim_{n\to\infty}a_nx=0 \right\}, $$则称$H$是$\u27e8mathcal{I}\u27e9$-特征化的。我们结合理想$\u27e8mathcal{I}\u27e9$的结构与拓扑性质，研究这类子群的描述集合论复杂度与可Polish化性。\n我们的核心结构结果表明：$\u27e8mathcal{I}\u27e9$是解析P-理想，当且仅当所有$\u27e8mathcal{I}\u27e9$-特征化子群都是可Polish化的。在该条件下，我们明确构造了一种相容的更细Polish群拓扑。借助可Polish化子群的相关结果，我们得到这类子群可能的Borel复杂度的三分性。\n若$\u27e8mathcal{I}\u27e9$是广义密度理想，我们证明了更精细的二分性：每个真$\u27e8mathcal{I}\u27e9$-特征化子群要么是可数的，要么是$F_{\sigma\delta}$-完全的。我们还通过构造一个由可和理想特征化的、既非$F_\sigma$也非$F_{\sigma\delta}$-完全的子群，证明该二分性对一般解析P-理想不成立。最后，我们明确描述了与幂序列、斐波那契序列、阶乘序列相关的子群，并以若干开放问题作结。
英文摘要
Given an ideal $\mathcal I$ on $ω$, a subgroup $H$ of the unit circle $\mathbb T$ is said to be $\mathcal I$-characterized if there exists a sequence of integers $(a_n:n\inω)$ such that $$ H= \left\{ x\in\mathbb T: \mathcal I\text{-}\lim_{n\to\infty}a_nx=0 \right\}. $$ We investigate the descriptive complexity and Polishability of these subgroups in terms of the structural and topological properties of the ideal $\mathcal I$. Our main structural result shows that $\mathcal{I}$ is an analytic $P$-ideal if and only if all $\mathcal I$-characterized subgroups are Polishable. In such case, we explicitly describe a compatible finer Polish group topology. Using results on Polishable subgroups, we obtain a trichotomy for their possible Borel complexities. If $\mathcal{I}$ is a generalized density ideal, we show the sharper dichotomy that every proper $\mathcal I$-characterized subgroup is either countable or $F_{σδ}$-complete. We also prove that this fails for general analytic $P$-ideals by constructing a subgroup characterized by a summable ideal which is neither $F_σ$ nor $F_{σδ}$-complete. Finally, we give explicit descriptions of the subgroups associated with the sequences of powers, the Fibonacci sequence, and the sequence of factorials. We conclude with several open questions.
