An approach to a problem in network design using genetic algorithms

This paper describes a new approach to finding solutions to the optimal communication spanning tree problem (OCSTP) using a genetic algorithm. The difficulties posed by this problem are reviewed and a genetic algorithm that consistently finds very good, if not optimal, solutions to it is presented. Finally, a comparison of the genetic algorithm's solutions to those produced by a good heuristic is given that demonstrates the genetic algorithm's ability to find solutions at least equivalent to, if not superior to, those found by the heuristic. Copyright © 1995 Wiley Periodicals, Inc., A Wiley Company