Higher \(K\)-theory of toric varieties (Q1417758)

The fundamental theorem of \(K\)-theory asserts that the higher \(K\)-theory of a regular ring \(R\) is homotopy invariant, that is, \(K_i(R)= K_i(R[T])\) for \(i\geq 0\), where \(R[T]\) is the polynomal ring. A natural way to generalise this result is to replace the polynomial ring \(R[T]\) by a monoid ring \(R[M]\), where the monoid \(M\) is commutative, cancellative, torsion-free, and without non-trivial units. It is known that the fundamental theorem does not hold in this generality. In its stead, the author considers a `nilpotence conjecture'. To state this, note that for any natural number \(c\) there is a homothety \(c: M\to M\), \(m\mapsto cm\) (writing \(M\) additively), which induces group homomorphisms \(c_*:K_i(R[M])\to K_i(R[M])\), \(i\geq 0\). Let \((c_1, c_2,\dots)\) be any sequence of natural numbers, each \(c_h\geq 2\). The conjecture is then that for any \(x\in K_i(R[M])\), there is some \(j= j(x)\) with \((c_1\cdots c_j)_*\in K_i(R)\). The author uses his geometric theory of monoids and the technique of pyramidal descent to show that the conjecture `almost' holds over a field \(k\), in that \((c_1\cdots c_j)_*\in \text{Im\,}K_i(\Lambda)\) for a certain ring \(\Lambda\) that is not too far removed from \(R\). In characteristic \(0\), this result, together with a further condition on the action of big Witt vectors on the \(K\)-theory of graded algebras, establishes the conjecture. Further, the author finds an explicit infinite series of polyhedral cones for which the conjecture holds when \(k\) is an algebraic number field.