Тип публикации: статья из журнала
Год издания: 2017
Идентификатор DOI: 10.1007/978-3-319-61833-3_6
Ключевые слова: Cooperation meta-heuristics, Linear time invariant system, Multi-objective genetic algorithm, multi-objective optimization, parameters identification
Аннотация: Cooperative multi-objective optimization tool is proposed for solving the order reduction problem for linear time invariant systems. Normally, the adequacy of an order reduction problem solution is estimated using two different criteria, but only one of them identifies the model. In this study, it was suggested to identify the paraПоказать полностьюmeters using both of the criteria, and since the criteria are complex and multi-extremum there is a need for a powerful optimization algorithm to be used. The proposed approach is based on the cooperation of heterogeneous algorithms implemented in the islands scheme and it has proved its efficiency in solving various multi-objective optimization problems. It allows us to receive a set of lower order models, which are non-dominated solutions for the given criteria and an estimation of the Pareto set. The results of this study are compared to the results of solving the same problems using various approaches and heuristic optimization tools and it is demonstrated that the set of solutions not only outperforms these approaches by the main criterion, but also provides good solutions with another criterion and a combination of them using the same computational resources.
Журнал: Lecture Notes in Computer Science (см. в книгах)
Выпуск журнала: Т.10386 LNCS
Номера страниц: 49-56
ISSN журнала: 03029743
Издатель: Springer-Verlag GmbH