The industrial manufacturing of poly(vinyl chloride) (PVC) by suspension polymerization is carried out in batch reactors. The productivity increase of these processes is directly related to the reduction of the time required to complete each batch. The reactor cooling system is designed such that it is capable to compensate for the maximum rate of heat release by the exothermic polymerization, so that the reactor cooling capacity is underutilized. An attractive mode of operation is to run the industrial process isothermally and to use a mixture ("cocktail") of different initiators, which is able to spread the polymerization rate over the batch time. In this work, the optimal formulation of an initiator mixture is studied as a dynamic optimization problem, which makes use of a representative mathematical model for the batch suspension vinyl chloride polymerization process. Results show that by choosing the optimal amounts of initiators in the cocktail, a significant reduction of the total processing time for a given polymer specification can be obtained, as compared to the case in which only one initiator is optimally chosen.