Immobile indices and CQ-free optimality criteria for linear copositive programming problems (Q2174923)

The article by O.I. Kostyukova, T.V. Tchemisova, and O.S. Dudina is a valuable contribution to so-called copositive programming, which means linear programming over the convex conic set of copositive matrices; scientists showed that many difficult, so-called NP-hard optimization problems can be expressed as copositive ones. This paper is rigorous in a mathematical and theoretically sense, and will certainly become relevant practically so that, based on it, further investigations can be made and real-world applications conducted. These may address and involve ``normal forms'' of problems, new numerical codes, and become fruitful in Operational Research, Economics, Finance, and time-dependent or switching problem classes like Game Theory and Optimal Control theory, such as Hybrid Systems, Stochastic Games and Stochastic Optimal Control. This paper aims to deeply investigate programs of linear ``copositive'' optimization in which the feasible sets just contain vectors for which the quadratic forms that are induced by corresponding linear matrix combinations regarded on the nonnegative orthant (generalized ``quarter space'') are nonnegative. Provided such a linear copositive program, the authors define for its constraints ``immobile indices'' and a ``normalized immobile index set''. They demonstrate that this set is either empty or representable as a union of finitely many convex closed and bounded polyhedrons. In fact, studying the structure of that set and associated properties of the feasible set provides new optimality criteria for those programs. Actually, the new criteria do not need any additional conditions (such as constraint qualifications). With an illustrative example the authors illustrate the novel optimality conditions and that they enable for a detection of optimality of feasible solutions, where known sufficient optimality conditions failed in this regard. The authors apply their new ``immobile indices'' approach to formulate new regularized primal and dual problems that are explicit and even ensure strong duality. This excellent paper is well-structured, deep, well written and well exemplified. The six sections of this paper are these: 1. Introduction, 2. Linear Copositive Programming Problem, 3. Optimality Conditions for Linear Copositive Problems, 4. Example, 5. Dual Formulations of Copositive Problems: the Standard Lagrangian Dual and the Regularized Dual Problems, and 6. Conclusions and the Future Work. In the future, extensions and refinements in analytic theory, e.g., with semi-infinite or disjunctive optimization, and computational techniques may be expected within the lively scientific family, inspired and prepared by this paper as well. These might be made in terms of in all areas of academia, of modern industries, societies and economies, in medicine, bio- and neuroscience, in geo- and earth- and environmental sciences, as well as for developing all our nations.