Hasse principle and weak approximation for multinorm equations (Q466133): Difference between revisions

Let \(L_i\) be finite separable extensions of a global field \(k\) and \(a\) an element of \(k^{\star}\). Consider \[ \prod_{i=1}^{n} N_{L_{i}/k}(z_i) = a \] with its affine \(k\)-variety \(X\). It is known [\textit{J. J. Sansuc}, J. Reine Angew. Math. 327, 12--80 (1981; Zbl 0468.14007)] that the obstructions to the Hasse principle and weak approximation for \(X\) are measured in terms of certain finite groups. The main result is theorem 6: With \(F:=\cap_{i=1}^{n} L_i\), let \[ N_{F/k}(w)=a \] with its affine \(k\)-variety \(Y\). Under some technical assumptions, it is proved that the finite groups measuring obstruction to Hasse principle and weak approximation associated to \(X\) and \(Y\) are isomorphic. The proof of the main result is first done for linearly disjoint field extensions, where there is no obstruction to Hasse principle and weak approximation. A first corollary is an analogue for weak approximation of a conjecture by \textit{T. P. Pollio} and \textit{A. S. Rapinchuk} [J. Number Theory 133, No. 2, 802--821 (2013; Zbl 1309.11078)] concerning the connection between the norm principle for the intersection of two finite Galois extensions of \(k\) and a multinorm principle. A second corollary proves a special case of the conjecture itself. However, a counterexample provided shows that the general case of the conjecture is false. There is a heuristic discussion involving the applicability of the fibration method of why the weak approximation analogue and a special case of the conjecture hold.