As most interior point algorithms iterate, they repeatedly perform costly matrix operations, such as projections, on the entire constraint matrix. For large-scale linear programming problems, such operations consume the great majority of the computation time required. However, for problems where...