Numerical semigroups with large embedding dimension satisfy Wilf's conjecture. (Q1939494)

Let \(S\) be a numerical semigroup, that is, a submonoid of the set of nonnegative integers, \(\mathbb N\), with finite complement in \(\mathbb N\). The Frobenius number of \(S\), \(F(S)\), is largest integer not belonging to \(S\). The conductor of \(S\), \(c(S)\), is defined as \(c(S)=F(S)+1\). Notice that \(c(S)+\mathbb N\subseteq S\). The least positive integer in \(S\) is the multiplicity of \(S\), denoted by \(m(S)\). Set \(n(S)\) to be the cardinality of the set \(\{s\in S\mid s<F(S)\}\). Let \(S^*\) be the set of nonzero elements of \(S\). The unique minimal generating system of \(S\) is \(S^*\setminus (S^*+S^*)\), and its cardinality, \(\nu(S)\), is the embedding dimension of \(S\). Wilf's conjecture states that \(c(S)\leq n(S)\nu(S)\). The papers by \textit{D. E. Dobbs} and \textit{G. L. Matthews}, [Int. J. Commut. Rings 2, No. 4, 195-202 (2003; Zbl 1165.13300)] and \textit{N. Kaplan}, [J. Pure Appl. Algebra 216, No. 5, 1016-1032 (2012; Zbl 1255.20054)] collect several classes of numerical semigroups fulfilling Wilf's conjecture. The author adds to this list a couple of families. The first one is the set of numerical semigroups with multiplicity less than or equal to twice the embedding dimension. In order to show that a semigroup \(S\) in this family satisfies Wilf's conjecture, the author proves that the conjecture holds if and only if \[ \sum_{k=0}^{L-1}(n_k(S)\nu(S)-m(S))+(n_L(S)\nu(S)-\rho)\geq 0, \] where \(L\) and \(\rho\) are positive integers such that \(c(S)=Lm(S)+\rho\), \(1\leq \rho\leq m\), and \(n_k(S)\) is the cardinality of the set \(\{s\in S\mid s<F(S),\;km(S)\leq s<(k+1)m(S)\}\). Then he describes the summands in terms of the Apéry set of the semigroup with respect to the multiplicity. As a consequence all numerical semigroups with multiplicity less than or equal to 8 satisfy Wilf's conjecture. The second class presented is that of numerical semigroups generated by a generalized arithmetic sequence \((m,hm,hm+2d,\dots,hm+ld)\), with \(\gcd(m,d)=1\), \(m\geq 2\), \(l\leq m-2\). In this setting the author uses that the type of a semigroup \(S\) in this family (the cardinality of the set of integers \(z\) such that \(z+S\subseteq S\), denoted \(t(S)\)) is less than the embedding dimension, and then the known fact that \(c(S)\leq n(S)(t(S)+1)\).