Concordance homomorphisms from knot Floer homology (Q2362595): Difference between revisions

Invariants coming from the Heegaard-Floer homology of a knot have been used to produce many powerful results in knot concordance. In the present paper appears a one-parameter family of ``\(t\)-modified Heegaard-Floer homologies''. A resulting numerical invariant \(\Upsilon_K(t)\) gives a homomorphism from \(\mathcal{C}\), the knot concordance group to the space of continuous piecewise linear functions on \([0,2]\). The pointwise properties of the derivative of this function are related to the \(\tau\)-invariant. This function is seen to be symmetric in that \(\Upsilon_K(t)= \Upsilon_K(2-t)\). The pointwise evaluation of this function gives bounds on the 4-genus of \(K\). The authors compute \(\Upsilon_K\) when \(K\) is quasi-alternating or is a torus knot. By doing so, they exhibit that the map \(K\mapsto \left(\frac{1}{n}\Upsilon_K\left(\frac{2}{n}\right)\right)\) gives an epimorphism from \(\mathcal C\) onto \(\mathbb Z^\infty\). Since an epimorphism from an abelian group onto \(\mathbb Z^\infty\) automatically splits, they find a \(\mathbb Z^\infty\)-summand of \(\mathcal C\) as an immediate consequence. In Section 7 of the paper they recall the formula for the Heegaard-Floer homology of a satellite knot coming from recent work of Lipshitz-Ozsváth-Thurston. In Section 8, they use it to compute \(\Upsilon_K\) for certain Whitehead doubles and their cables. In doing so they produce a \(\mathbb Z^\infty\)-summand of \(\mathcal C\) containing only topologically slice knots. The authors go on to compare their invariants with the \(\epsilon\)-invariant due to Hom and to generalize their construction to the setting of links.