The design challenge in multipurpose batch plants is optimizing plant configuration and equipment size under the constraints of multiple production plans, and these decision variables are interactive with each other through scheduling. To optimize these decision variables, the evolutionary design method on the basis of scheduling according to cyclic production (cyclic scheduling) is appropriate, and batch composition of a cycle (number of batches and their relative batch sizes per one-cycle) for respective production plans should be specified beforehand. In this study, a method to specify the batch composition that enables optimal design is developed. From the viewpoint of optimizing plant configuration, the common batch sizes for a multiple number of production plans (core batch sizes) are introduced for each product, and specifying the batch composition on the basis of these core batch sizes is converted into assignment problems, from the viewpoint of optimizing equipment sizes. These assignment problems are formulated into IP (Integer Programming) models, and the optimal design satisfying multiple production plans becomes possible. The effectiveness of this method is demonstrated through an example problem.