Shuffling algorithm for boxed plane partitions (Q1011405)

The authors consider discrete time Markov processes that preserve uniform measures on boxed plane partitions. A Markov one-step changes the size of the box from \(a\times b\times c\) to \((a-1)\times (b+1)\times c\) or \((a+1)\times (b-1)\times c\). Algorithmic realization of each step involves \(O((a+b) c)\) operations. An application is an efficient perfect random sampling algorithm for uniformly distributed boxed plane partitions.Trajectories of the Markov processes considered can be viewed as random point configurations in the three-dimensional lattice. The authors compute the bulk limits of the correlation functions of the resulting random point process on suitable two-dimensional sections. The limiting correlation functions define a two-dimensional determinantal point process with certain Gibbs properties.