The basis{bi}i=1d is Minkowski-reduced if bihas minimum length among all vectors in L linearly independent from{bj}j=1i−1. Equivalently, bihas minimum length among all vectors vsuch that {b1,…,bi−1,v}can be extended to form a basis ofL. Such a notion is strongest among all lattice reduction notions and is generally extremely hard to compute. Another equivalent definition is
∥bi∥≤j=i∑dcjbjgcd(cj)=1
Bounds
λi(L)2≤∥bi∥2≤max(1,(45)i−4)λi(L)2
The proof presented here is based off [Waerden 1956]. We proceed by induction. Letbibe a Minkowski-reduced basis for some latticeL. The lower bound is immediate and for i=1, the upper bound is immediate as well.
Let v1,v2…vibe linearly independent vectors such that∥vj∥=λj(L). Let Li−1be the sublattice generated by b1,b2,…bi−1. Evidently somekmust exist such thatvkis not in Li−1. Consider the new lattice L′=L∩span(b1,b2,…bi−1,vk). Letvk′be the shortest vector inL′−Li−1such thatb1,b2,…,bi−1,vk′is a basis for L′and we have
Ifn=1, then we are done sincevkcan be extended to a basis of L, so ∥bi∥≤∥vk∥=λk(L)≤λi(L). Otherwise, we have n2≥4. Let vk′=p+qwherepis the projection ofvk′inLi−1. Since by definition we have∥p∥2≤∥p±bi∥2, we must have
but since λi(L)2≤∥bi∥2by definition, the case of i=2,3cannot occur here (hence n=1in these cases).
Exercises
1) Show that both definitions of Minkowski-reduced lattice are equivalent
2) Consider the lattice L=2000102001002010002100001. We have showed in a previous exercise that the successive minima are all2but no basisbican satisfy ∥bi∥=λi, show that for any Minkowski reduced basis bi, the basis must satisfy ∥bi∥2=max(1,(45)i−4)λi(L)2