Benchmark Instances for Robust Combinatorial Optimization with Budgeted Uncertainty

Zenodo (CERN European Organization for Nuclear Research)

Abstract

We provide a set of instances for robust combinatorial optimization under budget uncertainty that have been described and used for benchmarking in the paper "A Branch & Bound Algorithm for Robust Binary Optimization with Budget Uncertainty", published in Mathematical Programming Computation by Christina Büsing, Timo Gersing and Arie Koster. To create this set of test instances, we use nominal instances from the MIPLIB 2017, which we transform into robust instances by defining robustness budgets Γ and deviations of the objective coefficients. In total, 67 instances from the collection of the MIPLIB 2017 are considered. For each of these, we provide data for constructing 12 corresponding robust instances by combining four different values for Γ with three different ranges for the deviation of the objective coefficients.

Authors 3

1. Timo Gersing Aachen

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University

2. Christina Büsing Aachen

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University

3. Arie M. C. A. Koster Aachen

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University