Analysis and Adjustment of a Genetic Algorithm for Non-Permutation Flowshops

The viability of many heuristic procedures strongly depends on the adequate adjustment of parameters. This work presents an adjustment procedure which was applied to a Genetic Algorithm. First, a preliminary analysis is performed, intended to obtain a better understanding of the behavior of the parameters, as for example to estimate how likely it is for the preceding adjustment of the parameters to remain in local minima. Special attention is paid on the variability of the solutions with respect to their repeatability. The four phases of the adjustment procedure are Rough-Adjustment, Repeatability, Clustering and Fine Adjustment.