This paper introduces a comprehensive extension to structured distance to instability for continuous-time linear systems, developing a unified framework for multiple perturbation structures and novel computational algorithms.
Key findings
Unified framework for structured distance to instability accommodating affine, block-diagonal, and sparsity-constrained perturbations.
Theoretical characterization of structured distance via structured singular values and spectral theory.
Novel subspace-based algorithms with local quadratic convergence.
Extensions to time-delay systems and mixed uncertainty structures.
Complete experimental validation methodology with benchmarks and metrics.
Limitations & open questions
Computing structured stability radii remains NP-hard, and scalable algorithms are limited.