Some new axioms for triangulated categories (Q755876)

The octahedral axiom has always been viewed as the ``mysterious'' axiom of triangulated categories. The thrust of this article is that the right object to study is not octahedra, but mapping cones of maps of triangles (thus, the author proposes that the right notion of a triangulated category is one that comes along with a category of triangles and ``good'' maps). So, a triangulated category (new version) is an additive category \({\mathcal T}\) with an automorphism \(\Sigma\) : \({\mathcal T}\mapsto {\mathcal T}\), together with a category \({\mathcal S}\) of triangles in \({\mathcal T}\). The author came across the problems studied in this article by trying to define a K-theory for triangulated categories. In the last part of this article he gives some applications of the results outside K- theory.