Skip to Main Content
LPSPS is a new linear programming program based on the saddle point algorithm which can solve large-scale, sparse linear programming problems. The saddle point algorithm is a new fast iterative method for linear programming. It has been widely used in many fields and shows better performance than other public domain simplex-based programs. Several aspects in the development of the program are described in this article such as the saddle point algorithm, pre-solving, post-solving and the data scaling. At the end of the paper, we compare LPSPS with CPLEX on the NETLIB test set.