Degenerate Optimal Lie Sectors in the Symmetrized Sedenion Khanhnion Model: Convex Faces, Directional Sensitivity, and Entropy Smoothing
Year: 2026
Type: preprint
Abstract: This paper studies least-squares Lie-sector fitting at repeated largest eigenvalues in the standard-real two-channel symmetrized sedenion Khanhnion model. Within the specified family of maximal rotation-invariant sectors, the optimal fits form a product of real projective spaces. A real density-matrix relaxation preserves the optimal value, and its optimal exposed faces consist precisely of densities supported on the top eigenspaces. Classical compressed-eigenvalue formulas yield one-sided directional derivatives of the optimal loss and characterize limiting directions of perturbed optimizers. An explicit covariance crossing demonstrates the impossibility of continuous exact single-sector selection through every degeneracy. Standard matrix-entropy regularization provides an analytic density-matrix surrogate, an additive value-error bound, and a zero-temperature limit equal to the normalized top-eigenspace projector. Mixed densities are explicitly distinguished from actual Lie-sector projectors. The manuscript includes proofs, Harvard-style references, a notation index, an AI-assistance disclosure, the author-provided emblem on the first page, and an explanatory diagram on the final page. No general spectral or entropy priority, bracket-preserving transport, or physical interpretation is claimed.
Author: Nguyễn Kim, Khánh
Institution: University of Khanh Hoa
Language: English
Topic: Matrix Theory and Algorithms
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7237274192
