The productivity in batch processes is related to reduction of the time required to complete each batch. An increase in productivity can be achieved by running the polymerization isothermally using a mixture of initiators with different decomposition rates. The amount of each initiator in the mixture can be optimized to increase productivity while not exceeding the maximum heat release that the cooling system is capable of compensating for. In this work, neural networks are used to search for the optimum operating condition of the reactor. The procedure to find the best neural network topology is presented as well as its application. (c) 2005 Wiley Periodicals, Inc.