System identification acceleration and improvement with genetic programming usage
Jana Nowaková, Jan Platoš, Martin Hasal · 2017
Identification of controlled system is a known and necessary task in proper controller design and its adaptation. In all cases, system identification is also a needed procedure for assessment of the nature of the systems and for the simulation proposes. Presented system identification is based on the difference equation parameters finding (optimization) with the usage of the genetics algorithms. The only disadvantage of the proposed method could be the computing time, which can be too long for real applications, when it is necessary to be really precise. One of the factor, which influences the computing time is the number of input-output data. The acceleration and improvement of the proposed method is done by the taking into the account the number of the data, which is necessary for the sufficiently precise identification of the system.