IMPLEMENTING DATA ENVELOPMENT ANALYSIS - PRIMAL OR DUAL

The computational performance of various implementations of a data env elopment analysis model are investigated. These include primal and dua l versions of both single-stage (epsilon represented as a small numeri cal magnitude) and two-stage (epsilon not explicitly represented) appr oaches. It is found, contrary to widely held belief, that the implemen tation based on the primal DEA model is faster than that based on the dual model, which has many fewer constraints. Some reasons for this un expected finding are suggested.