Sensitivity analysis on data envelopment analysis–DA model

Through the analysis of the feasible solution of the linear programming, the sensitivity analysis of data envelopment analysis–DA on linear assumption in the errors of sampled observations was discussed. Using the properties of base solution vectors of the linear restriction equations, a condition to maintain the same discriminant result between the sampled observations of changed and unchanged was obtained: the disturbed row vector of sampled observations was satisfied all linear equations whose coefficients were the base solution vectors of linear restriction equations. And some properties of discriminant functions were given.