![]() | Only 14 pages are availabe for public view |
Abstract In this chapter, five approaches based on swarm intelligence techniques for achieving cloud computing tasks scheduling has been proposed. MACOLB is used to find the near-optimal resource allocation for batch tasks in the dynamic cloud system and minimize the makespan of tasks. The proposed algorithm uses the same self-adapting criteria for the MACO control parameters but has an added load balancing factor. ACO, PSO and ABC for cloud task scheduling algorithms model the behavior ant colony, the behavior of Particle swarm and the behavior of honey bees respectively. Firstly the best values of parameters for each algorithm, experimentally determined. Then the algorithms in applications with the number of tasks varying from 100 to 1000 evaluated. Simulation results demonstrate that MACOLB algorithm outperforms the compared algorithms. MACO, ABC, PSO and ACO algorithms achieves better resource utilization and significantly outperforms FPLTF, random and FCFS algorithms on the basis of makespan and degree of imbalance. |