Open AccessGenetic algorithms - variable size populations of chromosomes. An adaptive approach
Author(s) -
R Maniu
Publication year - 2018
Publication title -
scientific bulletin
Language(s) - English
Resource type - Journals
eISSN - 2392-8956
pISSN - 1454-864X
DOI - 10.21279/1454-864x-18-i1-059
Subject(s) - population size , population , variable (mathematics) , chromosome , scroll , genetic algorithm , algorithm , acceleration , computer science , mathematics , mathematical optimization , biology , engineering , genetics , physics , mechanical engineering , mathematical analysis , demography , classical mechanics , sociology , gene
The size of the chromosome population is an essential parameter of genetic algorithms. A large population involves a large amount of calculations but provides a complete scroll of the search space and the increased probability of generating a global optimum. A small population size, through the small number of operations required, causes a quick run of the algorithm, with increasing the probability of detecting a local optimum to the detriment of the global one. This paper proposes the use of an adaptive, variable size of chromosome population. We will demonstrate that this approach leads to an acceleration of the algorithm operation, without having a negative impact on the quality of provided solutions.