Reviewing: Ayral et al. (2026) -"A quadratic Grassmann manifold optimization problem arising from quantum embedding methods" Reviewing: Ayral et al. (2026) -"A quadratic Grassmann manifold optimization problem arising from quantum embedding methods" 1
Syed Muntasir Mamun
Ayral et al. (2026) present the first comprehensive mathematical analysis and algorithmic framework for a non-convex quadratic optimization problem on the real Grassmann manifold (\operatorname{Gr}(m, \mathbb{R}^M)) that arises in generalized Density-Matrix Embedding Theory (DMET) for bath-orbital construction. The objective (J(P) = \operatorname{Tr}(BP) - \frac12 \operatorname{Tr}(APAP)) with (A \succeq 0) is shown to admit an Aufbau principle, first- and second-order optimality conditions, and a convex relaxation (\tilde{J}) over the spectrahedron whose gapped minimizer is Ayral, T., Cancès, E., Faulstich, F. M., Lin, L. and Negre, A. (2026) 'A quadratic Grassmann manifold optimization problem arising from quantum embedding methods', arXiv preprint arXiv:2603.17080v1 [math.OC]. 1 Mamun, SM (2026)…………………………………………1 Reviewing: Ayral et al. (2026) – "A quadratic Grassmann manifold optimization problem arising from quantum embedding methods" provably the global optimum of the original problem. Closed-form solutions are derived for commuting (A,B), perturbative regimes, and the DMET-specific block structure, yielding sharp necessary and sufficient conditions for full fragment-bath disentanglement that improve upon prior bounds. Also available on Academia.edu: https://www.academia.edu/attachments/132307859/download_file?s=portfolio
•Jan 1, 2026•0 citations