Open AccessFormation of initial point of initialization methods for optimization algorithms
Author(s) -
V. S. Tynchenko,
A. A. Pavlenko,
V. V. Bukhtoyarov,
Dmitriy Tikhonenko,
S. V. Tynchenko,
А. В. Цветцых
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1353/1/012115
Subject(s) - initialization , algorithm , sequence (biology) , computer science , point (geometry) , parametric statistics , global optimization , optimization problem , mathematical optimization , mathematics , statistics , genetics , geometry , biology , programming language
The initial point of initialization method is one of the main parameters for global optimization algorithms. Many scientists are engaged in its construction. The importance of this parameter for the entire algorithm is still not proven at all. Today, initialization methods based on stochastic algorithms are used. Six algorithms for constructing multidimensional points for global optimization algorithms – boolean strings – is designed. The available algorithm is analyzed. The authors use the starting points scattering algorithms, which are: LPτ sequence, UDC sequence, uniform random scatter. A large number of algorithms relaunches is used. The best way to initialize the starting points for the non-parametric, genetic algorithm, the MIVER scheme algorithm and the collective optimization method based on the Co-Operation of Biology Related Algorithms (COBRA) for these test functions has been determined.