Linear programming and the Newton barrier flow (Q1111465): Difference between revisions

The author investigates the interior algorithms for linear programming. These algorithms are connected with some trajectories of vectors approximating an optimal solution to a linear programming problem. The author shows that given a bounded polyhedral set P with nonempty interior, the logarithmic barrier function (with no objective components) induces a vector field of negative Newton directions which flows from the center of P to the solution of every possible linear program of P.