Approximating semidefinite packing problems

In this paper we define semidefinite packing programs and describe an algorithm to approximately solve these problems. Semidefinite packing programs arise in many applications such as semidefinite programming relaxations for combinatorial optimization problems, sparse principal component analysis, and sparse variance unfolding technique for dimension reduction. Our algorithm exploits the structural similarity between semidefinite packing programs and linear packing programs.


Accepted to SIAM J. Optimization.



View Approximating semidefinite packing problems