Title :
On optimal population size of genetic algorithms
Author :
Alander, Jarmo T.
Author_Institution :
Dept. of Comput. Sci., Helsinki Univ. of Technol., Espoo, Finland
Abstract :
A description is given of the results of experiments to find the optimum population size for genetic algorithms as a function of problem complexity. It seems that for moderate problem complexity the optimal population size for problems coded as bitstrings is approximately the length of the string in bits for sequential machines. This result is also consistent with earlier experimentation. In parallel architectures the optimal population size is larger than in the corresponding sequential cases, but the exact figures seem to be sensitive to implementation details.<>
Keywords :
computational complexity; genetic algorithms; parallel architectures; bitstrings; genetic algorithms; optimum population size; parallel architectures; problem complexity; sequential machines; Computer science; Distributed computing; Genetic algorithms; Information processing; Laboratories; Parallel architectures; Parallel processing; Problem-solving; Robot control; Robustness;
Conference_Titel :
CompEuro '92 . 'Computer Systems and Software Engineering',Proceedings.
Conference_Location :
The Hague, Netherlands
Print_ISBN :
0-8186-2760-3
DOI :
10.1109/CMPEUR.1992.218485
