A network design problem by a GA with linkage identification and recombination for overlapping building blocks

Abstract

Efficient mixing of building blocks is important for genetic algorithms and linkage identification methods that identify interdependent variables tightly linked to form a building block have been proposed. However, they have not been applied to real-world problems enough. In this chapter, we apply a genetic algorithm incorporating a linkage identification method called D5 and a crossover method called CDC to a network design problem to verify its performance and examine the applicability of linkage identification genetic algorithms. © 2008 Springer-Verlag Berlin Heidelberg.