This short communication complements the DEA model proposed by Lovell and Pastor (Eur. J. Oper. Res. 118 (1999), 46-51), by incorporating both positive and negative criteria in the model. As such, we propose a DEA model, known as pure DEA, using a directional distance function approach.

The introduction of a framework for optimal coalition formation using data envelopment analysis (DEA) methods is the focus of this paper. Simple examples illustrate how DEA is useful in formulating coalition models and deriving optimal solutions. In particular, the paper shows the relevance of the proposed framework in the context of analyzing how companies may reach decisions to acquire potential partners.