Decision problems in Group Technology (GT) based Flexible Manufacturing Systems (FMS) involve the identification of part families and then grouping the required machines into manufacturing cells to provide much needed flexibility in the manufacturing process as a whole. The criteria used for making these decisions in the past haveusually been the similarities of parts with respect to the machine needed. In some cases, material handling costs have also been incorporated in the objective function. Cell sizes are usually exogenously determined. A more complex problem arises when cell as well as cell composition have to be endogenously determined as part of model formulation and solution. A further complication arises when the number of machines of each type required for meeting the demand has to be determined simultaneously with the cell size and composition decisions with respect to relevant cost criteria. This paper develops an integer programming model to determine simultaneously, cell size, capacity selection and cell membership in a GT-based flexible manufacturing system.