-------------------------------------------------------------------
Title: Computational experiences in solving large scale SDPs 
       by Krylov subspace methods 

Abstract: 
We report our computational experiences on solving large 
scale SDPs via a primal-dual interior-point method (IPM) 
for which the linear system in each iteration is solved 
by a preconditioned Krylov subspace method such as the PCG method. 
Large scale SDPs often arise from the SDP relaxation of hard 
combinatorial problems, and even larger ones arise if the semidefinite 
matrix variable is also constrained to have non-negative elements. 
In the second half of the talk, we discuss the construction of 
preconditioners and related computational issues for the linear systems 
arising from an SDP where its matrix elements are also constrained
by a large number of linear inequalities.   
-----------------------------------------------------------------------