Analysis of evolutionary process in a lot sizing application

This paper seeks to determine the relationship between the best chromosome and its ancestors in genetic algorithms. The domain chosen is that of the classical lot sizing problem. A genetic algorithm software is developed to find the best lot size. We trace the best chromosome back for all generations to its ancestors and observe the evolutionary process. The preliminary results show that not all promising solutions led to the best chromosome. There were several good chromosomes repeating from generation to generation in the later part of the process and eventually leading to the best chromosome.