Optimizing performance with multiple responses using cross-evaluation and aggressive formulation in data envelopment analysis

An efficient optimization procedure is proposed for improving a product/process performance with multiple responses using two Data Envelopment Analysis (DEA) techniques, including the cross-evaluation and aggressive formulation techniques. The experiments generated in a Taguchi orthogonal array are considered Decision-Making Units (DMUs). The multiple responses are set inputs and/or outputs for all DMUs. Cross-evaluation and aggressive formulation techniques are employed to measure a DMU’s performance. The efficiency scores are then adopted to identify the combination of process factor levels that optimizes a product/process performance with multiple responses. Finally, the proposed procedure is illustrated by three case studies previously reported in the literature. The computational results show that the aggressive formulation technique is the most efficient in optimizing performance compared with the cross-efficiency technique, principal components analysis, and genetic algorithm methods. In conclusion, the advantages of the proposed optimization procedure may motivate practitioners to implement it in order to optimize a product/process with multiple responses in a wide range manufacturing applications.