Boltzmann samplers for v-balanced cycles

This paper is devoted to the random generation of a new operator which we call v-balanced colored cycle, a cycle of which the components are assigned different colors with frequencies restricted by a vector v. We propose an efficient approximate-size sampler (of linear expected time complexity) following the Boltzmann model introduced by Duchon et al. (2004) [9]. One of our motivations is to show that the absence of a decomposable specification, which form the basis of the Boltzmann model, can be circumvented by mixing Boltzmann samplers with other methods of random generation.