K-adaptability in stochastic optimization

Abstract We consider stochastic problems in which both the objective function and feasible set are affected by uncertainty. address these using a K -adaptability approach, solutions for given problem computed before uncertainty dissolves afterwards best of them can be chosen realized scenario. analyze complexity resulting from theoretical viewpoint, showing that, even case deterministic solved polynomial time, deciding if solution exists is $$\mathcal {NP}$$ NP -hard discrete probability distributions. Besides we prove that an approximation factor underlying carried over to our problem. Finally, present exact approaches including branch-and-price algorithm. An extensive computational analysis compares performances proposed algorithms on large randomly generated instances.