Pixton's double ramification cycle relations (Q680301)

Let \(\overline{\mathcal M}_{g,n}\) be the Deligne-Mumford moduli space of stable genus \(g\) curves with \(n\) marked points. For a tuple of integers \(A = (a_1, \ldots, a_n)\) whose sum is zero, the double ramification cycle \(R_{g,A}\) is a special class (in the sense of relative Gromov-Witten theory) in the Chow group \(A^g(\overline{\mathcal M}_{g,n})\) whose restriction to the moduli space \(\mathcal M_{g,n}\) of smooth pointed curves is the class of the locus of pointed curves \((C, x_1, \ldots, x_n)\) that admit a branched cover \(f: C \to \mathbb P^1\) such that the positive \(a_i\) describe the ramification type over \(0\) and the negative \(a_i\) describe the ramification type over \(\infty\). Eliashberg raised the question of giving an explicit description of the double ramification cycle \(R_{g,A}\). \textit{C. Faber} and \textit{R. Pandharipande} [J. Eur. Math. Soc. (JEMS) 7, No. 1, 13--49 (2005; Zbl 1084.14054)] proved that \(R_{g,A}\) lies in the tautological ring of \(\overline{\mathcal M}_{g,n}\). Explicit formulas for \(R_{g,A}\) were further discovered by \textit{R. Hain} [in: Handbook of moduli. Volume I. Somerville, MA: International Press; Beijing: Higher Education Press. 527--578 (2015; Zbl 1322.14049)], \textit{S. Grushevsky} and \textit{D. Zakharov} [Proc. Am. Math. Soc. 142, No. 12, 4053--4064 (2014; Zbl 1327.14132)] and \textit{R. Cavalieri} et al. [J. Pure Appl. Algebra 216, No. 4, 950--981 (2012; Zbl 1273.14053)] in the loci of curves of compact type and curves with rational tails, respectively. Recently, \textit{A. Pixton} [``Double ramification cycles and tautological relations on \(\overline{\mathcal M}_{g,n}\)'', available from the author] constructed an explicit tautological class of mixed degree \(\Omega_{g,A}\) on the entire moduli space \(\overline{\mathcal M}_{g,n}\), and he conjectured that (1) \([\Omega_{g,A}]_g = R_{g,A}\) and (2) \([\Omega_{g,A}]_d = 0\) for all \(d > g\), where \([\cdot]\) is the codimension-\(d\) part of a class. Part (1) of the conjecture was recently proved by \textit{F. Janda} et al. [Publ. Math., Inst. Hautes Étud. Sci. 125, 221--266 (2017; Zbl 1370.14029)]. The main result of this paper proves Part (2) of the conjecture. Moreover, the authors show that the tautological relations out of Pixton's conjecture are a consequence of Pixton's \(3\)-spin relations constructed in [\textit{R. Pandharipande} et al., J. Am. Math. Soc. 28, No. 1, 279--309 (2015; Zbl 1315.14037)].