Personal tools



Jump to: navigation, search

Conference proceedings article

Title Sparse Simplex for Compressed Sensing
Author José M. N. Vieira, Dorabella Santos, Huiyuan Wang
Booktitle Proceedings of RecPad2009
Address Aveiro, Portugal
Month October
Year 2009
Group (before 2015) Signal Processing Laboratory
Indexed by ISI Not known yet
Scope National

In this paper we propose a new stopping criteria for the Simplex algorithm when the solution is known to be sparse. The method is based on the evolution of the cost function. When the cost function stalls, the sparsity of the solution is checked and the optimality of the solution is verified by using the dual problem. Moreover, we will show that when the degenerate sparse solution is found, the cost of the dual Simplex problem is the same as the cost of the primal problem, ensuring in that way that the solution is optimal. The proposed method ensures that when the Simplex algorithm is applied in Compressed Sensing, it stops as soon as the sparse solution is found and that it is optimal.