Equivariant noncommutative motives (Q2404193)

This article considers {differential graded (dg) categories} \(\mathcal A\) over a basefield \(k\), which is a category enriched over dg \(k\)-vector spaces. To every (dg) \(k\)-algebra \(A\) there is an associated dg category with a single object, and to any quasiseparated \(k\)-scheme \(X\) there is the category of perfect complexes \(\text{perf}(X)\) which has a canonical dg enhancement \(\text{perf}_{dg}(X)\). A dg category \(\mathcal A\) with an action of a finite group \(G\) is denoted \(G\circlearrowright\mathcal A\) and is called a \(G\)-dg category, and the associated dg categories of \(G\)-equivariant objects are denoted \(\mathcal A^G\). By composition with the functor \(G\circlearrowright\mathcal A\mapsto\mathcal A^G\) the invariants of dg categories \(E\) are pushforward to invariants of \(G\)-dg categories \(E^G\). The overall goal of this article is to study all the invariants simultaneously, and this includes the development of a theory of \(G\)-equivariant noncommutative motives. A symmetric monoidal functor \(U^G:G-\mathsf{dgcat}_{sp}(k)\rightarrow\mathsf{NChow}^G(k)\) from smooth, proper \(G\)-dg categories to \(G\)-equivariant noncommutative Chow-motives is constructed, and this is initial among all the invariants \(E^G\). The morphisms of \(\mathsf{NChow}^G(k)\) are defined by terms of the \(G\)-equivarant Grothendieck group of certain triangulated categories of bimodules. Thus in particular, the endomorphism ring of the \(\otimes\)-unit \(U^G(G\circlearrowright 0k)\) is identified with the ring of representations \(R(G)\). \textit{I. A. Panin} [\(K\)-Theory 8, No. 6, 541--585 (1994; Zbl 0854.19002)] constructed a motivic category \(\mathcal C^G(k)\) mixing smooth projective \(G\)-schemes and separable algebras. In this category twisted projective homogeneous varieties can be studied. One main result of the present article is the construction of a fully faithful symmetric monoidal functor from \(\mathcal C^G(k)\) to \(\mathsf{NChow}^G(k)\). This construction generalize the results of Panin to the invariants \(E^G\). Using the theory of \textit{D. Edidin} and \textit{W. Graham} [Invent. Math. 131, No. 3, 595--644 (1998; Zbl 0940.14003)] on equivariant intersection theory, \textit{R. Laterveer} [Indag. Math., New Ser. 9, No. 2, 255--275 (1998; Zbl 0946.14007)], \textit{J. N. Iyer} and \textit{S. Müller-Stach} [Doc. Math., J. DMV 14, 1--18 (2009; Zbl 1167.14006)] extended the theory of Chow motives to the \(G\)-equivariant situation. In the present article, this theory is related to \(G\)-invariant motives, embedding the the commutative theory fully faithfully into the noncommutative theory. The \(G\)-equivariant analogue \(K_0\mathsf{Var}^G(k)\) of the Grothendieck ring of varieties can be studied by considering \(G\)-equivariant motivic measures. The author proves that for \(X\) a smooth projective \(G\)-variety, the assignment \(X\mapsto U^G(G\circlearrowright\mathsf{perf}_{dg}(X))\) gives a \(G\)-equivariant motivic measure \(\mu^G_{nc}:K_0\mathsf{Var}^G(k)\rightarrow K_0(\mathsf{NChow}^G(k))\) with values in the Grothendieck ring of the category of \(G\)-equivariant noncommutative Chow motives. A lot of information can be read out of the measure \(\mu_{nc}^G\). The author gives the following applications: Let \(X\) be a smooth projective \(G\)-scheme. One of the main applications in the article is a connection between the equivariant Chow motive \(\mathfrak h^G(X)_{\mathbb Q}\) and the \(G\)-category of perfect complexes \(G\circlearrowright\text{perf}(X)\). This results in the following theorem, stated verbatim: Theorem. If \(\text{perf}(X)\) admits a full exceptional collection \((\mathcal E_1,\dots,\mathcal E_n)\) of length \(n\) of \(G\)-invariant objects, i.e., \(\sigma^\ast(\mathcal E_i)\simeq\mathcal E_i\) for every \(\sigma\in G\), then there exists a choice of integers \(r_1,\dots,r_n\in\{0,\dots,\dim(X)\}\) such that \(\mathfrak h^G(X)_{\mathbb Q}\simeq\mathbb L^{\otimes r_1}\oplus\cdots\oplus \mathbb L^{\otimes r_n}\) where \(\mathbb L\in\text{Chow}(k)_{\mathbb Q}\) stands for the \(G\)-equivariant Lefschetz motive. This theorem can be applied to any \(G\)-action on a projective space, and it shows that the existence of a full exceptional collection of \(G\)-invariant objects indicates the \(G\)-equivariant Chow motive \(\mathfrak h^G(X)_{\mathbb Q}\). The sequence \(r_1,\dots,r_n\) above gives the unique indeterminacy which cannot be refined. However, the author remarks in a corollary that for every good \(G\)-cohomology theory \(H^\ast_G\), \(H^\ast_G(X)=0\) for \(i\) odd and \(\sum_i\dim H^i_G(X)=n\). This implies that the length of a hypothetical full exceptional collection of \(G\)-invariant objects is equal to \(\sum_i\dim H^i_G(X)\), and that if \(H^i_G(X)\ncong 0\) for some odd integer \(i\), then such a full exceptional collection cannot exist. Also notice that the theorem above shows that the \(G\)-equivariant Chow motive \(\mathfrak h^G(X)_{\mathbb Q}\) loses all information about the \(G\)-action on \(X\), while the \(G\)-equivariant noncommutative Chow motive \(U^G(G\circlearrowright\text{perf}_{dg}(X))\) keeps some information about the \(G\)-action. This article considers representation theory, or invariant theory, of group-actions over a field \(k\) in the motivic theory. It concentrates mainly on the nice case, where \(\text{char}(k)\nmid |G|\). The results are important and gives nice applications of the motivic representation theory together with a good bridging of the usual representation theory with this motivic, equivariant theory.