A Modified Linear Program for Columnar Methods in Mathematical Programming

This paper develops a modified linear programming algorithm for columnar methods in mathematical programming that is applicable to the decomposition principle of linear programming, to concave maximization problems over compact and convex sets, and to searching generalized Lagrange multipliers. It derives the linear program from linear inequalities in the multiplier space and compares it with the well-known linear program of Dantzig and Wolfe for solving such problems.