"Cartan decomposition of general linear groups"의 두 판 사이의 차이

수학노트
둘러보기로 가기 검색하러 가기
imported>Pythagoras0
imported>Pythagoras0
57번째 줄: 57번째 줄:
 
==computational resource==
 
==computational resource==
 
* https://drive.google.com/file/d/1LV3AvkCQAGB_ifEzpy8V-96mq-2a2amO/view
 
* https://drive.google.com/file/d/1LV3AvkCQAGB_ifEzpy8V-96mq-2a2amO/view
 +
[[분류:migrate]]

2020년 11월 13일 (금) 06:31 판

introduction

\( \newcommand{\pmat}[4]{\begin{pmatrix} #1 & #2 \\ #3 & #4\end{pmatrix}} \def\GL#1{\mathrm{GL}_{#1}} \newcommand{\Q}{\mathbb{Q}} \newcommand{\Z}{\mathbb{Z}} \newcommand{\Qp}{\Q_p} \newcommand{\Zp}{\Z_p} \newcommand{\HH}{\mathcal{H}} \newcommand{\fsph}{f_{\mathrm{sph}}} \)

application to Hecke operators

  • Let \(G = \GL2(\Qp)\) and \(K = \GL2(\Zp)\)
  • Cartan decomposition \[G = \bigcup_{(m,n)\in \Z^2 : m\geq n} K\pmat {p^m} 0 0 {p^n} K\]
  • The Hecke operator \(T_p\in \HH(G,K)\) is given by convolution with the characteristic function of \(K\pmat p 0 0 1 K\)
  • Similarly, the operator \(R_p\) is given by convolution with the characteristic function of \(K \pmat p 0 0 p K\)
  • How \(T_p\) and \(R_p\) act?
  • The double coset for \(T_p\) decomposes as

\[ K \pmat p 0 0 1 K = \bigcup_{b=0}^{p-1} \pmat p b 0 1 K \bigcup \pmat 1 0 0 p K . \]

  • Hence

\[ \begin{aligned} (T_p \fsph)(1) & = \int_{K}\sum_{b}^{p-1} \fsph\left(\pmat p b 0 1 g \right)+ \fsph\left(\pmat 1 0 0 p g \right)\, dg \\ & = \fsph\left(\pmat p b 0 1 g \right)+ \fsph\left(\pmat 1 0 0 p g \right) \\ & = p \chi_1(p)|p|^{1/2}+p \chi_2(p)|p|^{-1/2} \\ & = p^{1/2}(\chi_1(p)+\chi_2(p)). \end{aligned} \]

  • The double coset for \(R_p\) is the single coset \(\pmat p 0 0 p K\), so

\[ \begin{aligned} (R_p\fsph)(1) & = \int_K \fsph\left(\pmat p 0 0 p g \right)+ dg \\ & = \fsph\left(\pmat p 0 0 p g \right) \\ & = \chi_1(p)\chi_2(p). \end{aligned} \]

computational resource