Inexact DC algorithms in Hilbert spaces with applications to PDE-constrained optimization
Submitted to Applied Mathematics & Optimization (2026)
Abstract
We develop inexact adaptive algorithms for minimizing difference-of-convex (DC) functions in Hilbert spaces. The proposed method, I-ADCA, allows inexact subgradients and inexact solutions of the convex subproblems while keeping global convergence to stationary points, and it admits explicit convergence rates under a Polyak–Łojasiewicz-type property. The main application is elliptic optimal control with control constraints and a nonconvex \(L^1-L^2\) sparsity-promoting penalty; combined with finite element discretization, this gives an efficient solution procedure with proven well-posedness and error bounds, confirmed by numerical experiments.
Cite
P. D. Khanh, V. V. H. Khoa, B. S. Mordukhovich, D. B. Tran, and Nghia V. Vo. Inexact DC algorithms in Hilbert spaces with applications to PDE-constrained optimization. Submitted to Applied Mathematics & Optimization, 2026. arXiv:2601.06622.