Local decomposition methods for linear programming (Q1091939)

This paper deals with yet another method for solving a number of linear programming problems that are linked by common constraints. The Dantzig- Wolfe technique and Benders' method for solving this class of problems are well-known. Two methods are proposed, one for linear programming problems linked by common constraints, which is called dual local decomposition method, and one for problems linked by common variables, called the primal local decomposition method. The main feature of these methods is that parametric solutions to the subproblems are interacting with the principal problem. According to the author the efficiency of the method will depend to a large extent on the problem structure which will determine how many subproblems are included into the principal problem in the course of the iterations.