Averaging \(t\)-structures and extension closure of aisles (Q2350622): Difference between revisions

The notion of a \(t\)-structure was introduced by Beilinson, Bernstein and Deligne [\textit{A. A. Beilinson} et al., Astérisque 100, 172 p. (1982; Zbl 0536.14011)] in order to construct perverse sheaves. \(t\)-structures have since become a major tool in understading the structure of triangulated categories. A lot of different \(t\)-structures can be found in the same triangulated categories. The authors focus on the problem of studying when such \(t\)-structures can be `averaged' into one \(t\)-structure. More precisely, let \((X_i, Y_i)_{i \in I}\) be a finite set of \(t\)-structures in a triangulated category \(T\), then one can define two pairs of subcategories \((X^I, Y^I)\) and \((X_I, Y_I)\) by \[ X^I := \langle X_i | i \in I \rangle \qquad Y^I := \bigcap_{i \in I} Y_i \] \[ X_I := \bigcap_{i \in I} X_i \qquad Y_I := \langle Y_i | i \in I \rangle \] where \(\langle C \rangle\) denotes the smallest full subcategory of \(T\) containing \(C\) and closed under extensions and direct summands. The natural question that arises is the following: are \((X^I, Y^I)\) and \((X_I, Y_I)\) \(t\)-structures in \(T\)? The authors find the conditions for a positive answer to this question in different settings, namely triangulated categories with set-indexed coproducts and a possibly infinte set of compactly generated \(t\)-structures (Theorem A, section \(2\)), \(k\)-linear, Krull-Schmidt triangulated categories (Theorem B, section \(3\)) and \(k\)-linear, piecewise tame domestic hereditary triangulated categories (Theorem C, section \(6\)). It is worth to mention the very clear introduction to \(t\)-structures and to the piecewise hereditary triangulated categories. These parts, together with the well-written proof and the examples make the article very clear to read and also a source of inspiration for other works on this topic.