Supplier evaluation and selection problems are inherently multicriteria decision problems. Numerous analytical techniques ranging from simple weighted scoring to complex mathematical programming approaches have been proposed to solve these problems. However, traditional supplier selection models too often fail to consider the interaction and the capacity interdependency among the suppliers. Suppliers may exhibit internal interactions if the evaluation criteria used for one supplier are believed to be significantly affected by the evaluation criteria used by one or more of the other suppliers in the group. We propose a new branch-and-bound algorithm that generates portfolio alternatives based on Data Envelopment Analysis (DEA). The DEA model proposed in this study evaluates alternative supplier portfolios with a multicriteria model that considers possible interactions among the suppliers.