This paper presents and discusses different unified frameworks that can be used for design of multipurpose batch plants. The first framework is based on the state−task network (STN) process representation proposed by Kondili, Pantelides, and Sargent (1993, 1998) for the scheduling of batch plants. The second is the maximal state−task network (m-STN) where design and operational characteristics are represented simultaneously (Barbosa-Povoa and Machietto, 1994). Finally, the third framework uses the resource−task network (RTN), proposed by Pantelides (1994) for the scheduling problem. The application of these three representations to the design problem is explored and generalized formulations are developed, which are described and characterized. A set of examples is solved and results are compared and discussed, with an emphasis on the adequacy and effectiveness of these representations to the modeling and solution of the design of multipurpose batch plants.