Сибирский федеральный университет

Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: International Workshop “Hybrid methods of modeling and optimization in complex systems” (HMMOCS 2022); Krasnoyarsk; Krasnoyarsk

Год издания: 2022

Идентификатор DOI: 10.15405/epct.23021.40

Ключевые слова: Binary differential evolution, genetic algorithm, black-box optimization

Аннотация: Differential evolution is known as a simple and well-argued evolutionary algorithm that demonstrates the high performance in many hard black-box optimization problems with continuous variables. The main feature of differential evolution is the difference-based mutation. The mutation explores the search space using the distribution of points in the population and usually can well adapt to the objective function landscape. There exist some modifications of differential evolution for the binary search space. The proposed approach involves the understanding of the binary space topology for developing a better analogue of the difference-based mutation. We have compared the proposed binary differential evolution algorithm with the standard binary genetic algorithm using a set of binary test problems, including hard deceptive problems. The preliminary experimental results shows that new binary differential evolution is a competitive search algorithm and outperforms the binary genetic algorithm in reliability for some problems but yields it in the required number of function evaluations. The proposed approach for developing mutation in DE can be expanded to all other mutation schemes for increasing the performance.

Журнал: HYBRID METHODS OF MODELING AND OPTIMIZATION IN COMPLEX SYSTEMS

Номера страниц: 328-335

Место издания: London, United Kingdom

Издатель: European Proceedings

• Sopov Evgenii (Siberian Federal University)

• РИНЦ (eLIBRARY.RU)