Load balancing task scheduling based on Multi-Population Genetic Algorithm in cloud computing

In this paper, a Multi-Population Genetic Algorithm (MPGA) considering load balancing is adopted for solving task scheduling problems in cloud environment instead of Genetic Algorithm to avoid premature convergence. In order to boost the search efficiency, the min-min and max-min algorithm are used for the population initialization. Moreover, Metropolis criterion is used in this paper to screen the offspring so that poor individuals can also be accepted with a certain probability, then the population diversity can be maintained and the local optimum can also be avoided. The simulation results show that a better task scheduling result (shorter completion time, lower processing costs, load balancing) could be achieved through the MPGA-based task scheduling algorithm, which means the algorithm can realize an effective task scheduling and is more suitable for handling quantities of tasks compared to adaptive genetic algorithm (AGA).