On knot Floer homology and cabling (Q2571381): Difference between revisions

Let \(K\) be a knot in a solid torus \(T\) and \(e\colon T\to S^{3}\) be an embedding, where \(S^{3}\) is the three-sphere. Denote by \(C\) the core of \(e(T)\). Then the image \(e(K)\) is a knot in \(S^{3}\) and called a satellite of \(C\) with pattern \(K\). It is well known that the Alexander polynomial of the satellite \(e(K)\) is expressed in terms of the Alexander polynomials of \(K\) and \(C\) together with the winding number of \(K\) in \(T\). Let \(\widehat{H\!F\!K}(K,i)\) be the knot Floer homology introduced by \textit{Z.~Szabó} and \textit{P.~Ozsváth} [Adv. Math. 186, No.~1, 58--116 (2004; Zbl 1062.57019)]. There it was proved that the Alexander polynomial \(\Delta(K;t)\) can be determined by \(\widehat{H\!F\!K}(K,i)\), i.e., \(\sum_{i}\chi\left(\widehat{H\!F\!K}(K,i)\right)t^{i}=\Delta(K;t)\), where \(\chi\) denotes the Euler characteristic. So it is a natural question to ask how the knot Floer homology of a knot and its satellite are related. In the paper under review the author studies a cable knot, a satellite knot where the pattern is a torus knot. For a \((2,n)\)-cable of any knot \(K\) with large \(| n| \), the author shows that the knot Floer homology depends only on the filtered chain homotopy type of \(\widehat{C\!F\!K}(K)\).