We consider problems of linear copositive programming where feasible sets consist of vectors for which the quadratic forms induced by the corresponding linear matrix combinations are nonnegative over the nonnegative orthant. Given a linear copositive problem, we define immobile indices of its constraints and a normalized immobile index set. We prove that the normalized immobile index set is either empty or can be represented as a union of a finite number of convex closed bounded polyhedra. We show that the study of the structure of this set and the connected properties of the feasible set permits to obtain new optimality criteria for copositive problems. These criteria do not require the fulfillment of any additional conditions (constraint qualifications or other). An illustrative example shows that the optimality conditions formulated in the paper permit to detect the optimality of feasible solutions for which the known sufficient optimality conditions are not able to do this. We apply the approach based on the notion of immobile indices to obtain new formulations of regularized primal and dual problems which are explicit and guarantee strong duality. © 2020, Springer Nature B.V.

Authors

Journal

Language

English

Status

Published

Link

Year

2020

Organizations

^{1}Institute of Mathematics, National Academy of Sciences of Belarus, Surganov str. 11, Minsk, 220072, Belarus^{2}CIDMA–Center for Research and Development in Mathematics and Applications, Department of Mathematics, University of Aveiro, Campus Universitário Santiago, Aveiro, 3810-193, Portugal^{3}Belarusian State University, Independence Ave., 4, Minsk, Belarus^{4}RUDN University, 6, Miklukho-Maklaya st., Moscow, 117198, Russian Federation

Keywords

Constraint qualification; Copositive programming; Normalized immobile index set; Optimality conditions; Semi-infinite programming; Strong duality

Date of creation

10.02.2020

Date of change

10.02.2020

Share

Urologiia (Moscow, Russia : 1999).
2020.
P. 125-130

Communications in Computer and Information Science.
Springer Verlag.
Vol. 1145 CCIS.
2020.
P. 56-71