Characterization of Approximate Efficient Solutions of Nonconvex Vector Optimization Problems
摘要
In this paper, we develop new conic scalarization theorems to characterize approximate proper efficiency relative to improvement sets in a real locally convex space. By leveraging seminorm-linear separation functions and Bishop-Phelps-type separating cones introduced by Günther, Khazayel, and Tammer [SIAM J. Optim., 34 (2024), pp. 225-250], we establish and rigorously prove these results. Notably, our theorems extend Kasimbeyli’s foundational conic scalarization framework – originally formulated for normed spaces (Kasimbeyli [SIAM J. Optim., 20 (2010), pp. 1591-1619]) – to the broader context of real locally convex spaces.