Mixed alternative in the assignment by classes vs. conventional methods for calculation of individuals lifetime in GAVaPS

The paper presents a novel method referred to as assignment by classes for the calculation of the lifetime of individuals in genetic algorithms with varying population size (GAVaPS). The proposed strategy groups individuals according to their fitness. The purpose is to use the allowed range of lifetime values in a way which is more suitable than conventional strategies (proportional, linear and bilinear) to search for the optimum. Three possibilities of the assignment by classes were implemented, but only one of them is analyzed in detail, the one called mixed alternative. A comparative study between this alternative and the conventional methods is carried out. Results are shown over 5 functions. Finally some conclusions are presented, along with possible future lines of work.