The equivariant coarse Novikov conjecture and coarse embedding (Q2216214)

This paper proves the equivariant coarse Novikov conjecture for certain group actions on certain geometric spaces. The method in this paper has its origin in a paper of Guoliang Yu [\textit{G.-L. Yu}, Invent. Math. 139, No.~1, 201--240 (2000; Zbl 0956.19004)]. The equivariant coarse Novikov conjecture claims that \[ \operatorname{Ind}_{\Gamma}:\displaystyle\lim_{d\to\infty}K^\Gamma_*(P_d(X)) \to K_*(C^*(X)^\Gamma) \] is injective. By Yu's method, one introduces the equivariant localization algebra and puts the index map in the following commutative diagram where the index map \(\operatorname{Ind}_{\Gamma,L}\) is an isomorphism under some mild conditions on \(X\): \[ \begin{tikzcd} {} & \lim\limits_{d\to\infty}K_*(C^*_L(P_d(X))^\Gamma) \dar["e_\ast" ']\\ \lim\limits_{d\to\infty}K_*^\Gamma(P_d(X)) \urar["\operatorname{Ind}_{\Gamma,L}"] \urar[phantom, "\circlearrowright", very near end, yshift=-1.0em, xshift=1em] \rar["\operatorname{Ind}_\Gamma" '] & K_*(C^*(P_d(X))^\Gamma)\,. \end{tikzcd} \] In this paper, \(X\) is a proper metric space with bounded geometry, \(\Gamma\) is a countable discrete group acting on \(X\) properly and isometrically. A subset \(Y\) of \(X\) is a fundamental domain for the \(\Gamma\)-action if \(X=\bigcup_{\gamma\in\Gamma}\gamma Y\). \(X\) has bounded distortion which means that for any \(\gamma\in\Gamma\), \(\sup_{x\in Y}d(\gamma x,x)<\infty,\) where \(Y\) is a fundamental domain for the \(\Gamma\)-action. Also, \(X/\Gamma\) and \(\Gamma\) both admit coarse embeddings into a Hilbert space. Under the condition that \(X/\Gamma\) admits a coarse embedding into a Hilbert space, the authors introduce an equivariant \(\pi\)-localization algebra \(C^*_{\pi,L}(X)^\Gamma\), where \(\pi:X\to X/\Gamma\), and prove that \[ (e_\pi)_*:\lim_{d\to\infty}K_*(C^*_{\pi,L}(P_d(X))^\Gamma)\to\lim_{d\to\infty}K_*(C^*(P_d(X))^\Gamma) \] is injective. Then the rest of the paper is devoted to prove that, for the inclusion map \(\tilde{\tau}:C^*_L(P_d(X))^\Gamma\to C^*_{\pi,L}(P_d(X))^\Gamma\), the induced map \[ \tilde{\tau}_*:K_*(C^*_L(P_d(X))^\Gamma)\to K_*(C^*_{\pi,L}(P_d(X))^\Gamma) \] is injective under the assumption that \(X\) has bounded distortion and \(\Gamma\) admits a coarse embedding into Hilbert space. Putting all these conclusions into one commutative diagram, one has \[ \begin{tikzcd} {} & \lim\limits_{d\to\infty}K_*(C^*_L(P_d(X))^\Gamma)\dar["e_*"]\rar["\widetilde{\tau}_*"] & \lim\limits_{d\to\infty}K_*(C^*_{\pi,L}(P_d(X))^\Gamma)\dlar ["(e_\pi)_*" ] \dlar [phantom, "\circlearrowright", very near end, yshift=1.4em ]\\ \lim\limits_{d\to\infty}K_*^\Gamma(P_d(X)) \urar["\operatorname{Ind}_{\Gamma,L}"]\urar[phantom, "\circlearrowright" ', yshift = -0.9em, xshift=1em, very near end] \rar["\operatorname{Ind}_\Gamma"] *& K_*\left(C^*(X)^\Gamma\right) & {}\\ \end{tikzcd} \] Therefore, \(\operatorname{Ind}_\Gamma=(e_\pi)_\ast\circ\tilde{\tau}_\ast\circ \operatorname{Ind}_{\Gamma,L}\) is injective.