Extension of primal-dual interior point methods to diff-convex problems on symmetric cones

被引:3
|
作者
Valkonen, Tuomo [1 ,2 ]
机构
[1] Univ Jyvaskyla, Dept Math Informat Technol, SF-40351 Jyvaskyla, Finland
[2] Karl Franzens Univ Graz, Inst Math & Sci Comp, Graz, Austria
关键词
symmetric cone; Jordan algebra; diff-convex; filter method; interior point; JORDAN ALGEBRAS; NONCONVEX OPTIMIZATION; ALGORITHMS; SUM;
D O I
10.1080/02331934.2011.585465
中图分类号
C93 [管理学]; O22 [运筹学];
学科分类号
070105 ; 12 ; 1201 ; 1202 ; 120202 ;
摘要
We consider the extension of primal dual interior point methods for linear programming on symmetric cones, to a wider class of problems that includes approximate necessary optimality conditions for functions expressible as the difference of two convex functions of a special form. Our analysis applies the Jordan-algebraic approach to symmetric cones. As the basic method is local, we apply the idea of the filter method for a globalization strategy.
引用
收藏
页码:345 / 377
页数:33
相关论文
共 50 条