Heavy-Tailed Mutation with a Regularly Varying Constraint on the Distribution Function of Its Rate - Доклады на конференциях | Sciact - Система учета научной деятельности ИМ СО РАН

Heavy-Tailed Mutation with a Regularly Varying Constraint on the Distribution Function of Its Rate Доклады на конференциях

Язык

Английский

Тип доклада

Секционный

Конференция

XXIII International Conference Mathematical Optimization Theory and Operations Research
30 июн. - 6 июл. 2024 , Омск

Авторы

Eremeev Anton ^1,3 , Topchii Valentin ^1,2

Организации

1	Омский филиал ФГБУН «Институт математики им. С.Л. Соболева СО РАН».
2	Новосибирский государственный университет
3	Омский государственный университет им. Ф.М. Достоевского

We study the heavy-tailed mutation operator proposed by Doerr, Le, Makhmara, and Nguyen (GECCO 2017). The power-law assumption of mutation rate is generalizeed using regularly varying constraint on the distribution function of mutation rate. It is shown that, on the OneMax function class, the expected runtime of evolutionary algorithms with this generalized version of mutation is linear in the problem dimension.

Eremeev A. , Topchii V.
Heavy-Tailed Mutation with a Regularly Varying Constraint on the Distribution Function of Its Rate
XXIII International Conference Mathematical Optimization Theory and Operations Research 30 Jun - 6 Jul 2024