Napredna pretraga

Pregled bibliografske jedinice broj: 16055

Adaptive Genetic Algorithm


Jakobović, Domagoj; Golub, Marin
Adaptive Genetic Algorithm // Proceedings of the 20th International Conference on Information Technology Interfaces / Kalpić, Damir ; Hljuz Dobrić, Vesna (ur.).
Pula: SRCE University Computing Centre, 1998. str. 519-524 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)


Naslov
Adaptive Genetic Algorithm

Autori
Jakobović, Domagoj ; Golub, Marin

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of the 20th International Conference on Information Technology Interfaces / Kalpić, Damir ; Hljuz Dobrić, Vesna - Pula : SRCE University Computing Centre, 1998, 519-524

Skup
20th Int. Conf. on Information Technology Interfaces, ITI '98

Mjesto i datum
Pula, Hrvatska, 16-19.06.1998

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Genetic algorithm; local and global optima; adaptive genetic operators

Sažetak
In this paper we introduce an adaptive, 'self-contained' genetic algorithm (GA) with steady-state selection. This variant of GA utilizes empirically based methods for calculating its control parameters. The adaptive algorithm estimates the percent of the population to be replaced with new individuals (generation gap). It chooses the solutions for crossover and varies the number of mutations, all regarding the current population state. The state of the population is evaluated by observing some of its characteristic values, such as the best and worst individual's cost function (fitness) values, the population average etc. Furthermore, a non-uniform mutation operator is introduced, which increases the algorithm's efficiency. Adaptive method does not, however, restrict the applicability in any way. The described GA is applied to optimization of several multimodal functions with various degrees of complexity, employed earlier for comparative studies. Some deceptive problems were also taken into consideration, and a comparison between the adaptive and standard genetic algorithm has been made.

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo



POVEZANOST RADA


Projekt / tema
036014

Ustanove
Fakultet elektrotehnike i računarstva, Zagreb