Postoptimality for mean-risk stochastic mixed-integer programs and its application (Q1935908)

The mean-risk stochastic mixed-integer programming problem is considered where the recourse cost vector, the technology matrix, and the right-hand side vector in the second-stage problem are all random. The contamination technique is adopted for the postoptimality analysis. The continuity and differentiability of the optimal value function with respect to the parameter in the contaminated distribution is proved. The obtained results are applied to investigate the postoptimality of a stochastic mixed-integer programming problem with risk objective where the objective nonlinearly depends on the probability distribution.