This paper examines a scheduling problem from a fine chemicals company that synthesizes active pharmaceutical ingredients. The main goal is to select from the set of available equipment units, those units that are better-suited for the production of the ingredient being considered. A minimum number of them. maximizes plant flexibility by leaving some free capacity to respond faster to other product orders, but also increases the makespan, because significant changeover times are mandatory. A tradeoff is thus involved that is quantified by an objective function that minimizes a weighted sum of the total equipment allocation cost plus the makespan, which is penalized with a cost factor. Because several product batches are needed, a stable mode of operation is assumed, with most of the schedule being derived by repetition of an optimal production pattern. This is obtained through the use of a resource-task network discrete-time periodic scheduling formulation that generates a mixed-integer linear program (MILP). For this particular case study, it is shown that the periodic schedule contains all the necessary information for easily deriving the startup and shutdown phases of production. The results have shown that the industrial problem is tractable and that a few minutes of computational time are sufficient to determine the optimal cycle time and corresponding schedule. To allow for more-complex problems to be tackled, a new decomposition strategy is proposed that reduces the computational effort by one order of magnitude without severely compromising optimality.