View : 642 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.advisor김선영-
dc.contributor.author김현정-
dc.creator김현정-
dc.date.accessioned2016-08-26T03:08:46Z-
dc.date.available2016-08-26T03:08:46Z-
dc.date.issued2013-
dc.identifier.otherOAK-000000075692-
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/205078-
dc.identifier.urihttp://dcollection.ewha.ac.kr/jsp/common/DcLoOrgPer.jsp?sItemId=000000075692-
dc.description.abstractA class of quadratic optimization problems (QOPs) can be formulated in completely positive cone programs as shown in Arima, Kim and Kojima. In this thesis, solution methods for solving the class of QOPs are studied. Using the results for QOPs whose coefficient matrices have nonpositive off-diagonal (ODN-Assumption), we find a new class of completely positive programming problems that can be solved exactly by semidefinite programming and second order cone programming relaxations. We note that the boundedness of variables is very important to guarantee ODN-Assumption.;Arima, Kim 그리고 Kojima에 의해 2차 최적화 문제의 특정한 부류가 CPP(completely positive cone programming)로 변형이 가능하다는 것이 알려졌다. 이 논문은 그 부류에 속한 문제들을 푸는 방법에 대해 연구하였다. 모든 계수 행렬의 비대각항들이 양수가 아닌 2차 최적화 문제의 결과들을 사용하여, SDP 완화 방법(semidefinite programming relaxation)과 SOCP 완화 방법(second order cone programming relaxation)으로 CPP 문제를 풀 수 있도록 하는 새로운 부류의 CPP 문제를 찾을 것이다. 여기서 주목해야 할 점은 계수 행렬의 비대각항들이 양수가 아님을 보장하기 위해 가능 해들의 집합이 유계여야 한다는 것이다.-
dc.description.tableofcontents1. Introduction 2. Quadratic Optimization Problems 2.1 A general form of QOP and its SDP and SOCP relaxations 2.2 A class of QOPs solved by SOCP relaxation and its exact solutions 3. Completely Positive Cone Programming 3.1 A basic form of QOP and its CPP relaxation 3.2 A class of QOPs solved by CPP relaxation 3.2.1 A bounded case 3.2.2 An unbounded case under weaker conditions 4. Main Results 4.1 A standard form of CPPs and its SOCP relaxation 4.2 A class of CPPs solved by SOCP relaxation 4.2.1 A general case that can be solved exactly 4.2.2 A general case that can be solved approximately 4.2.3 A homogeneous case as a special case 5. Numerical Examples 5.1 A standard quadratic problem 5.2 The MAX-CUT problem 6. Concluding Remarks References 국문초록-
dc.formatapplication/pdf-
dc.format.extent960558 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.subject.ddc500-
dc.titleSolving Quadratic Optimization Problems Formulated in Completely Positive Cone Programming-
dc.typeMaster's Thesis-
dc.format.pageiii, 33 p.-
dc.identifier.thesisdegreeMaster-
dc.identifier.major대학원 수학과-
dc.date.awarded2013. 2-
Appears in Collections:
일반대학원 > 수학과 > Theses_Master
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE