Upsilon-type concordance invariants (Q2292402)

In this paper, a generalization of the knot concordance invariant \(\Upsilon\) is constructed; as an application, it is shown that for odd \(q\geq 7\), the pretzel knot \(P(-2,3,q)\) is not concordant to a sum of algebraic knots. Let us give some background and details. Associating to a knot \(K\) its Floer chain complex \(\text{CFK}^{\infty}(K)\) up to stable equivalence \(\sim\) gives a homomorphism from the smooth concordance group to the group \(\mathcal{CFK}/\sim\) of stable equivalence classes of certain chain complexes (with tensor product as group operation) [\textit{J. Hom}, J. Knot Theory Ramifications 26, No. 2, Article ID 1740015, 24 p. (2017; Zbl 1360.57002)]. This concordance invariant is powerful; but extracting accessible information from it is non-trivial. One way to do so is given by the \(\Upsilon\)-invariant [\textit{P. Ozsváth} et al., Adv. Math. 315, 366--426 (2017; Zbl 1383.57020)], which associates to a class in \(\mathcal{CFK}/\sim\), represented by a chain complex \(A\), (and thus to a knot concordance class) a piecewise-linear function \([0,1]\to\mathbb{R}\) in the following way: at \(t\in [0,1]\), mix the two filtrations of \(A\) linearly (in a way determined by \(t\)) to obtain another filtration; let \(\Upsilon_t(K)\) be the minimal degree with respect to that filtration of a cycle of \(A\) that is not a boundary. In the paper at hand, the two filtrations are mixed more flexibly, in a way determined by a south-west region \(C\) (a subset of \(\mathbb{R}^2\) that is closed under walking towards the south or west). For every such \(C\), this yield a concordance invariant \(\Upsilon^C(K) \in\mathbb{R}\). Since \(\Upsilon_t = \Upsilon^C\) for \(C\) a certain south-west region whose boundary is a line with slope determined by \(t\), the construction of the paper is indeed more general than the original \(\Upsilon\) invariant. In fact, it also contains the \(h_i\) invariants constructed in Rasmussen's PhD thesis as a special case. Furthermore, Kim-Livingston's secondary \(\Upsilon\)-invariants are generalized [\textit{S.-G. Kim} and \textit{C. Livingston}, Q. J. Math. 69, No. 3, 799--813 (2018; Zbl 1402.57016)]. As an application, it is shown that the knot \(T_{8,5}\# -T_{6,5} \# -T_{4,3}\) is not concordant to a Floer thin knot. This cannot be detected using the original \(\Upsilon\)-invariant.