Тези доповіді до ІV Всеукраїнській Інтернет-конференції здобувачів вищої освіти і молодих учених «Інформаційні технології: теорія і практика»

The main problem solved was to combine the above-described modifications based on the idea of a guided local search algorithm [3], which is a meta-heuristic method that builds penalties during the search and uses them to help local search algorithms to get out of the local minimum. When the local search algorithm reaches the local minimum, this algorithm modifies the objective function in a certain way and the algorithm then works with this modified objective function, which is built to get out from the local optimum. A key issue in the guided local search method is how to modify the objective function, which is well handled by the approach proposed in HGDN.In general, the presented algorithm can be successfully applied to solve global optimization problems with different types of complexities, such as multimodality, non-separability, nonlinearity, non-differentiability, gully trap, the high dimension of search space, high computational complexity, and more