No label defined (Q4356360)

Let \(X\) be an infinite set and let \(X\choose n\) denote the set of \(n\)-element subsets of \(X\). Let \(V_n\) denote the space of rational valued functions on \(X\choose n\). Let \(A=\bigoplus_{n\geq 0} V_n\). Then \(A\) is an algebra via \(fg(X)=\sum f(Y)g(X\setminus Y)\) where \(f\in V_n\), \(g\in V_m\) and \(Y\) ranges over all subsets of size \(n\) in \(X\) (which has cardinality \(n+m\)). This is the reduced incidence algebra of the poset of finite subsets on \(X\) introduced by Rota. It is a commutative associative algebra but elements with finite support are nilpotent. If \(G\) is a permutation group on \(X\), let \(A^G\) be the subring of \(G\)-invariant functions. If \(G\) is oligomorphic, then \(A^G\) is finite dimensional in each degree. If \(G\) has a finite orbit, then \(A^G\) contains no nontrivial nilpotent elements. The author conjectures the converse and indeed considers the stronger conjecture. Let \(e\in V_1\) denote the constant function \(1\). Then \(e\) is a nonzero divisor in \(A\). Cameron conjectures that if \(G\) has no finite orbits, then \(e\) is prime in \(A^G\) -- this would imply that \(A^G\) is an integral domain. The author discusses some special cases where \(A^G\) is a polynomial algebra. He also considers the relationship between \(A^G\) and \(A^H\) where \(G\) is the stabilizer of a point \(x\in X\) and \(H\) is transitive on \(X\). For the entire collection see [Zbl 0867.00021].