الفهرس 
abstractOnly 14 pages are availabe for public view
 |  | from | 124 |  |  |
| | | from | 124 | | |
Abstract
Nowadays, IT services are provided and delivered over the Internet on demand for anyone in anywhere at any time so the Cloud Computing is the most recent computing paradigm. However, the technology is still not fully developed. There are still some areas that are needed to be focused on. The task scheduling in the Cloud Computing environment has a lot of awareness as the applications tasks are mapped to the available resources to achieve better results. Although, there are some existed task scheduling algorithms for Cloud environment, none of them have considered both of the computing cost (processor cost), and the communication cost (cost between processors) with the time of execution in the same time. PSO (Particle Swarm Optimization) algorithm and GA (Genetic Algorithm) are two approaches that can be used for task scheduling in the Cloud computing environment. According to these algorithms, application tasks are scheduled for the available resources to reduce either the execution cost or the execution time. In this thesis, a PSO algorithm and GA algorithm have been developed for scheduling tasks of workflow application where there are dependencies between tasks. These modified algorithms are used for allocating the tasks to the available resources to minimize the execution time in addition to the computation cost on the Cloud Computing environments with considering a bounded number of processors. The main principle of the modified algorithm MPSO (Modified Particle Swarm Optimization) or MGA (Modified Genetic Algorithm) is dividing the workflow application into tasks, these tasks are modeled by DAG (Directed Acyclic Graph) and tasks are sorted in each level according to their dependencies in a descending order for execution priority. Then, PSO or GA algorithm is used to calculate the basic operations to schedule tasks on the available processors using a selection strategy of scheduling heuristics for scheduling first the ready tasks. This selection of tasks will enable the modified algorithms to generate high-quality scheduling in the Cloud Computing environment. The MPSO and MGA algorithms have been evaluated using different combination of the cost and the time fitness functions (i.e., And, Sequence, and Best-To-Best operations) to improve the performance in terms of the speedup, resource utilization and the system efficiency. The results from the combinations of each algorithm are compared with each other and it is noted that the MPSO algorithm outperforms the MGA algorithm as MGA algorithm achieves better results for small data sizes but MPSO algorithm achieves better results for larger data sizes. Also, it is found that MPSO algorithm in the combination of least time to the least cost is the best combination from the five combinations of the two algorithms in minimizing both time and cost of the scheduling and maximizing performance measures in the Cloud Computing environment.