Public Bidding in Government Procurement - a New Methodology for Ranking Proposals Using Full Dimensional Efficient Facets of DEA Frontier

This work proposes a methodology to find appropriate weights to be assigned to attributes to evaluate bids for government procurements. The new methodology complies with the European Union procurement directives and provides transparency about the weights’ tradeoffs for attributes in government procurement. The theoretical foundation is based on a Data Envelopment Analysis (DEA) frontier composed of full dimensional efficient facets (FDEFs) as the framework for measuring performance and ranking proposals. It intends to reduce the discretion of judgments inherent to weight restrictions in DEA models, where experts are required to develop additional assumptions to avoid null weights assigned to criteria. It applies a non-radial non-oriented performance measurement which is a modification of the Enhanced Russel Graph Measure (ERGM) to search for the closest target. The methodology combines the use of FDEFs and ERGM in two phases. In the first phase, a mixed integer linear program is formulated and solved to find all FDEFs using a reference data set. These facets are then used in the second phase to evaluate the submitted proposals through the modified ERGM. The insertion of the DEA model in the operational environment is facilitated using concept maps. These allow structuring the problem and facilitating the engagement of agents, aiming at the validation of the results. We apply the new methodology to a realistic problem from the Brazilian navy, with the support of concept mapping to direct the process