Title: The solution space structure of planted constraint satisfaction problems with growing domains

Abstract: Planting a solution into the random RB model, which is a prototype of random constraint satisfaction problem (CSP) with growing domains, can generate very hard satisfiable CSP benchmarks. We study the solution space structure of the planted RB model. With constraint density growing, we find that this model goes through four phase transitions. In the replica symmetric phase, what we call the independent phase transition occurs, after which the planted cluster (cluster containing the planted solution) is separated from the giant cluster. Then the solutions except that in the planted cluster go through the same clustering phase transition and the same satisfiability phase transition as the random RB model. The planted cluster goes through the isolated phase transition, after which the planted cluster contains only one solution. This phase diagram provides strong evidence that this model can generate very hard satisfiable CSP benchmarks. For over constraint instances (where the constraint density is very large), we find that the configuration space has only a single energy valley, which makes the instances tractable. Experiments using Belief Propagation confirm the locations of the clustering, satisfiability (by configurations outside the planted cluster), and isolated phase transition points.