Study of different optimization strategies for analogue circuits

• Alexander Zemliak ^[1] ; Fernando Reyes ^[1] ; Sergio Vergara ^[1]
  1. [1] Benemérita Universidad Autónoma de Puebla

     Benemérita Universidad Autónoma de Puebla

     México

     
• Localización: Compel: International journal for computation and mathematics in electrical and electronic engineering, ISSN 0332-1649, Vol. 35, Nº 3, 2016, págs. 927-942
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • Purpose – The purpose of this paper is to propose further development of the generalized methodology for analogue circuit optimization. This methodology is based on optimal control theory. This approach generates many different circuit optimization strategies. The authors lead the problem of minimizing the CPU time needed for circuit optimization to the classical problem of minimizing a functional in optimal control theory.

    Design/methodology/approach – The process of analogue circuit optimization is defined mathematically as a controllable dynamical system. In this context, the authors can formulate the problem of minimizing the CPU time as the minimization problem of a transitional process of a dynamical system. To analyse the properties of such a system, the authors propose to use the concept of the Lyapunov function of a dynamical system. This function allows us to analyse the stability of the optimization trajectories and to predict the CPU time for circuit optimization by analysing the characteristics of the initial part of the process.

    Findings – The authors present numerical results that show that we can compare the CPU time for different circuit optimization strategies by analysing the behaviour of a special function. The authors establish that, for any optimization strategy, there is a correlation between the behaviour of this function and the CPU time that corresponds to that strategy.

    Originality/value – The analysis shows that Lyapunov function of optimization process and its time derivative can be informative sources for searching a strategy, which has minimal processor time expense. This permits to predict the best optimization strategy by analysing only initial part of the optimization process.