Problema de autovalores, otimização de funções matriciais e robustez de sistemas dinâmicos: uma abordagem algoritmica

Abstract

In this work we analyse three problems. In the first, we present some algorithms to solve the eigenvector and eigenvalue problems of the symmetric matrix, In lhe second we analyse the optimization problems with matricial constraints and finally in the third some robustness properties of linear continuons time dynamic systems are studied. The eigenvector and etgenvalue problems are two methods: the decomposition technique on the matrix and the optimization of the Rayleigh quotient. In the propose two algorithms based on the Newton method gradient method. To solve a class of lhe optimization problems with matricial constraints we propose a methodologie based on the cutting plane technique. Two exemples are treated. Finally, for a given linear continuons time we determine sufflcient conditlons for the existence feedback matrix such that the closed-Ioop system is sense that the pertubed system is asymptocally stable, analysed using of the system, second one we and conjugated dynamic systems of a constant robust in the sense that the perturbed system is asymptocally stable.