One of the influential components in the probabilistic seismic-hazard analysis (PSHA) is selecting appropriate ground motion prediction equations (GMPEs). To handle the source of epistemic uncertainty, the most common approach is considering several GMPEs in a logic tree framework. In such an approach, the selected GMPEs and the weights assigned for each one are usually determined by experts, mostly according to gut feeling. In order to reduce the role of subjectivity in selecting GMPEs, some statistical criteria based on goodness-of-fit measures were developed in recent years. In this study, in order to incorporate different criteria in the ranking procedure, we present to use data envelopment analysis (DEA) as a multi-criteria efficiency evaluation framework. One of the advantages of the DEA procedure is that it is able to determine the efficiency or inefficiency of each GMPE. Also, the DEA methodology is able to handle the situation in which several target intensity measures such as spectral amplitude in different structural periods are interested to be considered in the ranking procedure. In this study, the DEA procedure is applied to the Central Iran seismotectonic province. The results show that the DEA can well evaluate the efficiency of GMPEs, and can be used as supplements to existing ranking procedures.
