Network Data Envelopment Analysis with Intermediate Products Structure

Traditional Data envelopment analysis models make no hypothesis concerning the internal operations in a "black box". To open the "black box", we put forward a type of network DEA structure with several sub-DMUs linked by intermediate products. The input node and output node also represent the entrance and exit of the box. However, the input is assignable in the paper. We suggested a suitable model to deal with intermediate products between divisions with assignable inputs in the network structure. The model enables us to pry into the internal operations of DEA by network structure with intermediate products and assignable input. We finally illustrate the calculation procedure of the proposed algorithm by a numerical example.