ON ARTIN'S CONJECTURE, II: PAIRS OF ADDITIVE FORMS

From MaRDI portal





Let \(n\) and \(k\) be integers greater than 1. The subject of this paper is non-trivial \(p\)-adic solutions of pairs of additive equations of the form \[ \sum_{j=1}^n a_jx_j^k= \sum_{j=1}^n b_jx_j^k=0, \] with integer coefficients \(a_j\), \(b_j\). In the spirit of a conjecture of Artin, it is expected that if \(n>2k^2\), then the above pair of equations always has a non-trivial \(p\)-adic solution for every prime \(p\). When \(k\) is odd, this was indeed proved by Davenport and Lewis, but for even \(k\), the picture is less complete. For example, in Part I of this series of papers [Bull. Lond. Math. Soc. 31, 305-313 (1999; Zbl 0933.11015)], the authors proved that \(n\geq 6k^2\) is a sufficient condition unless \(k\) is a power of \(2\), and the second author showed that \(n>16k^2-26k\) suffices in the latter case [Additive forms of degree \(K=2^l\), J. Number Theory 46, 391-408 (1994; Zbl 0797.11039)]. The paper under review substantially advances research on this problem for even \(k\). Here it is established that, for any given integers \(a_j\), \(b_j\), the above pair of equations has a non-trivial \(p\)-adic solution, provided that one of the following three conditions holds; (i) \(n>2k^2\), \(p\) is an odd prime, and \(k\neq p^{\tau}(p-1)\) for natural numbers \(\tau\), (ii) \(n\geq 4k^2\) and \(p\) odd, (iii) \(n\geq\max\{8k^2/k_0, 2k^2+1\}\) and \(p=2\), where we write \(k=2^{\tau}k_0\) with \(\tau\geq 0\) and \(k_0\) odd. Therefore, the authors reach the expected restriction \(n>2k^2\), except only for the cases where \(k\) takes the shape \(k=p^{\tau}(p-1)\) for a prime \(p\) and a natural number \(\tau\), or \(k=3\cdot 2^{\tau}\). The proof uses Olson's theorem concerning finite \(p\)-groups, as in Part I of this series, in order to solve a related congruence. Non-singularity of a solution is considered by colouring the variables according to values of \(a_jb_j^{-1}\) modulo \(p\), and this new idea is a key to this significant improvement.











This page was built for publication: ON ARTIN'S CONJECTURE, II: PAIRS OF ADDITIVE FORMS

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3152216)