Redução de modelos lineares em tempo continuo
AUTOR(ES)
Reinaldo Giusti Egas
DATA DE PUBLICAÇÃO
2004
RESUMO
In this work the model reduction problem is revisited and formulated through convex programming constrained by linear matrix inequalities. The H2 and H°° norms are used as comparison criteria between the original and reduced order models, having as starting point the present continuous-time filtering results. A rank constraint in some variables, suitably developed, results in partially observable state space equations, generating a transfer function of reduced order that approximates the original system. This method is validated by comparisons between the results obtained herein and the ones provided by the well known balanced truncation procedure. Such comparison is done by using statistically generated systems. Finally, two flexible structures are studied, leading to valid approximations for systems of high order derived from the method developed here.
ASSUNTO(S)
sistema lineares programação (matematica)
ACESSO AO ARTIGO
http://libdigi.unicamp.br/document/?code=vtls000334979Documentos Relacionados
- Otimização dinâmica em tempo real utilizando modelos não-lineares simplificados
- Aplicação de modelos de tempo-contínuo para escolha de portfólio ótimo
- Redução do tempo de convergência em modelos baseados no naming game
- Redução de cromo hexavalente em filtro biológico de fluxo contínuo
- Um curso em modelos lineares