A conjugate gradient algorithm for sparse linear inequalities (Q915387): Difference between revisions

The author presents a conjugate gradient type method for solving a system of linear inequalities, which takes full advantage of the sparsity the problem may have. It is shown by numerical comparisons with the successive overrelaxation algorithm that this approach is attractive when a system has a relative large number of inequalities. The method works on two phases. Phase 1 either finds the minimum-norm solution of the system or detects the inconsistency. In phase 2 an approximate least-squares solution to the system is obtained.