컨텐츠 시작
학술대회/행사
초록검색
| 제출번호(No.) | 0126 |
|---|---|
| 분류(Section) | Contributed Talk |
| 분과(Session) | (AM) Applied Mathematics(including AI, Data Science) (AM) |
| 발표시간(Time) | 20th-D-13:20 -- 13:40 |
| 영문제목 (Title(Eng.)) |
A new full-newton step infeasible interior-point method for linear optimization |
| 저자(Author(s)) |
Jongkyu Lee1, Gyeong-Mi Cho2 Sungkyunkwan University1, Dongseo University2 |
| 초록본문(Abstract) | Many studies on interior-point methods have utilized kernel functions to find new search directions. Notably, classical kernel functions well-known in this context include self-concordant, self-regular, and eligible kernel functions. In this talk, we introduce a new class of kernel functions that differs from these classical counterparts. The newly defined class of kernel functions is much easier to check than the conditions required for classical kernel functions. Based on this new class of kernel functions, we propose a unified approach for a small-update full-Newton step infeasible interior-point method for linear optimization. Furthermore, we demonstrate that it has the best-known worst-case computational complexity in this methodology. |
| 분류기호 (MSC number(s)) |
90C05, 90C51 |
| 키워드(Keyword(s)) | linear optimization, infeasible interior-point method, kernel function, polynomial complexity |
| 강연 형태 (Language of Session (Talk)) |
Korean |