In this paper, we introduce and study a two-stage distributionally robust mixed binary problem (TSDR-MBP) where the random parameters follow the worst-case distribution belonging to an uncertainty set of probability distributions. We present a decomposition algorithm, which utilizes distribution separation procedure and parametric cuts within Benders' algorithm or Lshaped method, to solve TSDR-MBPs with binary variables in the first stage and mixed binary programs in the second stage. We refer to this algorithm as distributionally robust integer (DRI) L-shaped algorithm. Using similar decomposition framework, we provide another algorithm to solve TSDR linear problem where both stages have only continuous variables. We investigate conditions and the families of ambiguity set for which our algorithms are finitely convergent. We present two examples of ambiguity set, defined using moment matching, or Kantorovich-Rubinstein distance (Wasserstein metric), which satisfy the foregoing conditions. We also present a cutting surface algorithm to solve TSDR-MBPs. We computationally evaluate the performance of the DRI Lshaped algorithm and the cutting surface algorithm in solving distributionally robust versions of a few instances from the Stochastic Integer Programming Library, in particular stochastic server location and stochastic multiple binary knapsack problem instances. We also discuss the usefulness of incorporating partial distribution information in two-stage stochastic optimization problems.
We study two-stage stochastic mixed integer programs (TSS-MIPs) with integer variables in the second stage. We show that under suitable conditions, the second stage MIPs can be convexified by adding parametric cuts a priori. As special cases, we extend the results of Miller and Wolsey (Math Program 98(1):73-88, 2003) to TSS-MIPs. Furthermore, we consider second stage programs that are generalizations of the well-known mixing (and continuous mixing) set, or certain piecewise-linear convex objective integer programs. These results allow us to relax the integrality restrictions on the second stage integer variables without effecting the integrality of the optimal solution of the TSS-MIP. We also use four variants of the two-stage stochastic capacitated lot-sizing problems as test problems for computational experiments, and present tight second-stage formulations for these problems. Our computational results show that adding parametric inequalities that a priori convexify the second stage formulation significantly reduces the total solution time taken to solve these problems.
We present a homogeneous algorithm equipped with a modified potential function for the monotone complementarity problem. We show that this potential function is reduced by at least a constant amount if a scaled Lipschitz condition is satisfied. A practical algorithm based on this potential function is implemented in a software package named iOptimize. The implementation in iOptimize maintains global linear polynomial-time convergence properties while achieving practical performance. When compared with a mature software package MOSEK (barrier solver version 6.0.0.106), iOptimize solves convex quadratic programming problems, convex quadratically constrained quadratic programming problems, and general convex programming problems in fewer iterations. Moreover, several problems for which MOSEK fails are solved to optimality. We also find that iOptimize seems to detect infeasibility more reliably than general nonlinear solvers Ipopt (version 3.9.2) and Knitro (version 8.0).
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.