A Characterization of Simultaneous Optimization, Majorization, and (Bi-)Submodular Polyhedra

Martijn H. H. Schoot Uiterkamp
DOI: https://doi.org/10.1287/moor.2023.0054
IF: 2.215
2024-02-21
Mathematics of Operations Research
Abstract:Motivated by resource allocation problems (RAPs) in power management applications, we investigate the existence of solutions to optimization problems that simultaneously minimize the class of Schur-convex functions, also called least-majorized elements. For this, we introduce a generalization of majorization and least-majorized elements, called (a, b)-majorization and least (a, b)-majorized elements, and characterize the feasible sets of problems that have such elements in terms of base and (bi-)submodular polyhedra. Hereby, we also obtain new characterizations of these polyhedra that extend classical characterizations in terms of optimal greedy algorithms from the 1970s. We discuss the implications of our results for RAPs in power management applications and derive a new characterization of convex cooperative games and new properties of optimal estimators of specific regularized regression problems. In general, our results highlight the combinatorial nature of simultaneously optimizing solutions and provide a theoretical explanation for why such solutions generally do not exist.
mathematics, applied,operations research & management science
What problem does this paper attempt to address?