The single-machine scheduling problem with periodic maintenances and sequence-dependent setup times aims at scheduling jobs on a single machine in which periodic maintenances and setups are required. The objective is the minimization of the makespan. We propose an exact algorithm based on the iterative solution of three alternative arc-time-indexed models. Extensive computational experiments are carried out on 420 benchmark instances with up to 50 jobs, and on 405 newly proposed instances involving up to 150 jobs. We compare the results found by all formulations with those obtained by the best available mathematical formulation. All instances from the existing dataset are solved to optimality for the first time.