Ill-posedness with respect to the solvability in linear optimization (Q2496624)

The authors study extensively the notion of ill-posedness in linear optimization. The aim of the research is to characterize the linear optimization problems that are ill-posed, which is defined as those where arbitrarily small perturbations of the coefficients produce very different types of problems, e.g., bounded/unbounded or feasible/infeasible. The article begins with an overview of useful definitions common in linear optimization. The authors then define the notion of stability in linear problems and some preliminary properties are presented. The bulk of this very interesting article contains a series of theorems and lemmas with proofs, which aim to present an insight on certain conditions that may cause a particular linear problem to be ill-posed. The article finally concludes with a section on the link of the current work with related research on condition numbers associated with optimization problems.