This work considers a parallel algorithm for solving the multidimensional multiextremal optimization problems. The algorithm combines nested optimization scheme and Peano-type space filling curves for the dimensionality reduction. To decrease the number of iterations of the global algorithm, we use one of local tuning techniques based on the adaptive estimation of the global optimizer. Parallel algorithm with mixed local-global strategy of search is proposed as well. The efficiency of the parallel algorithm was investigated using a supercomputer. The speedup of the algorithm using several cores as compared with the serial algorithm has been demonstrated experimentally.
Skip Nav Destination
Research Article| October 20 2016
Local tuning in multilevel scheme of parallel global optimization
AIP Conf. Proc. 1776, 060006 (2016)
Konstantin Barkalov, Ilya Lebedev; Local tuning in multilevel scheme of parallel global optimization. AIP Conf. Proc. 20 October 2016; 1776 (1): 060006. https://doi.org/10.1063/1.4965340
Download citation file: