Geometric progressions in sumsets over finite fields (Q2463452)

By using some simple lemmata as e.g., lemma 4: Given \(m+1>1\) vectors in a \(d\)th dimensional vector space \(V\) over \(\mathbb{F}_q.\) There are at most \(mq^{d^2-1}\) square \(d\) by \(d\) matrices \(M\) over \(\mathbb{F}_q\) with \[ 0 = \sum_{0}^{m} v_i M^i. \] and also some substantial computations using character sums the authors get new results on the existence of long geometric progressions (orbit of a vector under the \(M^i\)'s) in a sumset \(A+B\) of \(V\) with large enough subsets \(A\) and \(B.\) More precisely their main result reads: Let \(k\) be upper bounded by \((q-1)^{1/2}.\) Then the smallest \(g\) such that for arbitrary subsets \(A,B\) of \(V\) with \(\text{card}(A) \text{card}(B) \geq g\) the sumset \(A+B\) contains a geometric progression of length \(k\) satisfy: \[ g \leq 20 q^{2d-2/k}. \] Some appropriate generalizations are also undertaken.