A GRASP for Computing Approximate Solutions for the Three-Index Assignment Problem

In this talk a greedy randomized adaptive search procedure (GRASP) is presented for computing approximate solutions to the NP-hard three-index assignment problem (AP3). A FORTRAN implementation of this GRASP is tested on several problem instances of the AP3. A parallelization strategy that combines GRASP and path relinking is also proposed and tested for the problem. Computational results indicate that the GRASP and the parallel hybrid heuristic provide good approximate solutions to a variety of AP3 instances.