Sensitivity Analysis For Mathematical Programs With Generalized Equation Constraints

In this thesis,we construct a mathematical programs with generalized equation constraints,and conduct sensitivity analysis on two perturbation forms of the problem,obtaining the upper estimate of the subdifferential of the value function for the perturbation problem with respect to the multipliers satisfying the Fritz John optimality condition.Then,we conduct sensitivity analysis on the parameterized mathematical programs with complementarity constraints(MPCC),and provide the upper estimate of the subdifferential of the value function for the parametric MPCC with respect to the set of M multiplier.Finally,we extend the problem model to the matrix case,that is,there are semidefinite cone complementary constraints(SDCMPCC)in the constraint condition.We perform sensitivity analysis of the value function for parametric SDCMPCC,which is a new result.When the local optimum satisfies the restricted inf-compactness and there is no nonzero abnormal multiplier constraint qualification(NNAMCQ),we obtain the upper estimate of the subdifferential of the value function for parametric SDCPMCC with respect to M multiplier.