A study of permutation operators for minimum span frequency assignment using an order based representation

The genetic algorithm (GA) described in this paper breeds permutations of t ransmitters For minimum span frequency assignment. The approach hybridizes a GA with a greedy algorithm, and employs a technique called Generalized Sa turation Degree to seed the initial population. Several permutation operato rs From the GA literature are compared. and results indicate that position based operators are more appropriate for this kind of problem than are orde r based operators. My offspring versus mid-parent correlation studies on cr ossovers show Pearson's correlation coefficient to be a reliable predictor of performance in most cases. Results presented herein represent improvemen ts over previously published results.