Two new algorithms for the linear assignment problem

Abstract

The linear assignment problem (AP) being among the first linear programming problems to be studied extensively,, is a fundamental problem in combinatorial optimization and network flow theory. AP arises in numerous applications of assigning personnel to jobs, assigning facilities to locations, sequencing jobs, scheduling flights, project planning and a variety of other practica.1 problems in logistics planning. In this thesis work, we seek for new approaches for solving the linear assignment problem. The main concern is to develop solution methods that exhibit some sort of parallelism. We present two new approaches for solving the assignment problem : A dual-feasible signature guided forest algorithm and a criss-cross like algorithm.