Research On Energy Optimization Algorithm Based On Task Scheduling In Cloud-Environment

With the widespread popularity of the Internet and the vigorous development of information network technology, the development of cloud computing is becoming mature. At the sametime, cloud computing has covered more and more application fields. Therefore, it’s significant to find a reasonable method of allocating resources for task scheduling to reduce energy consumption and the enterprise cost.The data center of cloud computing has many characteristics different from those of traditional data center, such as heterogeneity, open and dynamic, owing to its virtualization feature. At present, the data, center’s scale and the user community of the cloud computing are gradually expanding. It becomes more important to design an efficient scheduling policy of allocating resources for users to complete task request with shortest time and least cost.Based on the fact that tasks scheduling is an NP complete problem, this thesis firstly changes the Task scheduling into a knapsack problem with the "backpack" theory, to make full use of the computing resources and to avoid resource contention. Then, this article proposes a hybrid genetic algorithm GGA (Greedy and Genetic Algorithms) based on Greedy algorithm and Genetic algorithm GGA algorithm proposes a scheme to modify individual species considering the actual situation of assigned tasks, and, it sets up reasonable target function and fitness function.Secondly, this thesis designs a target function considering scheduling time and dependability for independent tasks in virtual environment, which is used as fitness function of Genetic algorithm. The specific scheduling algorithm combines the thoughts of Min- Min and Genetic algorithm that is why we call it MGA(Min-Min and Genetic Algorithms). This algorithm uses the result of Min-Min algorithm as initial solution, then the genetic algorithm is run on that base.Finally, the corresponding simulation experiments are designed for the above two task scheduling strategies. The experimental results show that the two scheduling policies have reached the desired result.