Evaluation of Testing Software Program Based on DEA with Fuzzy Window

Abstract

Based on the DEA (Data Envelopment Analysis) traditional model, an improved evaluation model is introduced for the fuzzy index value and fuzzy preference weight information. The improved model is applied in evaluation of automatic testing software programs. Firstly, the subjective index values are transformed into fuzzy numbers. Secondly, the subjective preference weights are constructed as a constrain of fuzzy window. Finally, the improved fuzzy DEA model is established to access the comprehensive evaluation of automatic testing software programs. A typical solving algorithm for the evaluation model is practiced. The instance simulation carries out the evaluation with direct subjective weight preference information, resulting in automatic testing software programs keeping consistence for different confidence levels.