Neural network-based heuristic algorithms for hypergraph coloring problems with applications
Journal of Parallel and Distributed Computing - Special section best papers from the 2002 international parallel and distributed processing symposium
| Hi-index | 0.00 |
The Set Covering Problem (SCP) is a well known combinatorial optimization problem, which is NP-hard. We conducted a comparative study of eight different approximation algorithms for the SCP, including several greedy variants, fractional relaxations, randomized algorithms and a neural network algorithm. The algorithms were tested on a set of random-generated problems with up to 500 rows and 5000 columns, and on two sets of problems originating in combinatorial questions with up to 28160 rows and 11264 columns. On the random problems and on one set of combinatorial problems, the best algorithm among those we tested was the neural network algorithm, with greedy variants very close in second and third place. On the other set of combinatorial problems, the best algorithm was a greedy variant and the neural network performed quite poorly. The other algorithms we tested were always inferior to the ones mentioned above.