In control theory the poles of a LTI system is computed from the eigenvalues of matrix pencil (A,E) where matrices A and E come from the descriptor system representation. In my problem I have a stable system (i.e. the all the eigenvalues have negative real part). Now if matrix A is perturbed by adding a all-ones matrix, almost all the poles stays same (i.e. stable), however some poles (maximum 2) become unstable (i.e. positive real part) and stay on the real axis very far from the origin. I have to mathematically prove the reason behind that. Can anyone help me answering that question please or at least direct me to some references where I can search for answers?  

More Muhammad Kabir's questions See All
Similar questions and discussions