The effects of two replacement strategies on a genetic algorithm for scheduling jobs on computational grids

Computational Grid (CG) represents a new computational framework whose efficient use requires schedulers that allocate user's tasks to the grid resources in an acceptable amount of time. In this paper, we study the effects of two replacement strategies on a GA for Job Scheduling on Computational Grids, namely Steady-State GA (SSGA) and Struggle GA (SGA). Considering the makespan, the experimental results show the improvement obtained by SGA over SSGA for moderate size instances. However, the time needed by the SGA to reach makespan values obtained by the SSGA rapidly increases as more jobs and machines are added to the CG.