A shortest augmenting path algorithm for dense and sparse linear assignment problems

Abstract

We develop a shortest augmenting path algorithm for the linear assignment problem. It contains new initialization routines and a special implementation of Dijkstra's shortest path method. For both dense and sparse problems computational experiments show this algorithm to be uniformly faster than the best algorithms from the literature. A Pascal implementation is presented. © 1987 Springer-Verlag.