Robust eigenvalue-clustering against J-structured uncertainty

The problem of matrix eigenvalue clustering in an EMI-region is handled. The considered matrices are subject to an additive J-structured uncertainty i.e. an additive term of the form JE where J is known and E unknown. The eigenvalues are robustly clustered if the 2-norm of E is less than a robustness bound. This proposed bound is derived owing to LMI tools. It reveals itself as very easy to compute and not too conservative.