Scheduling Algorithm Based On Genetic Algorithm Workflow Tasks. Cloud Environment

In recent years, with the development of cloud computing, a new mode for delivering and using resources, which is based on the model of producer-consumer and the principle of payment per use, has emerged. Because of its good benefits, more and more enterprises and research institutes have participated and used this commercial calculation and service mode widely. Similarly,workflow has also been widely used in scientific and other computing areas. In order to take advantage of cloud computing in resource consuming, and also in order to lower the cost of the investment, it makes good sense to put workflow tasks on cloud computing environment for execution. Under this service pattern, what the user concerns most is the processing time and execution cost of the task. Given that the existing task scheduling strategies only pay attention to the execution time without considering execution expenses, we have designed a workflow task scheduling algorithm based on both execution cost and processing time in the cloud computing environment in this paper, to obtain good user experience and economic benefits which both are of great importance.In this paper, we considered the factors about customer's service quality requirements and designed a genetic algorithm based on two scheduling heuristic respectively to produce satisfying customer's demand scheduling strategy, namely, to minimize the execution cost in a given deadline and the task processing time in a limited execution budget. Finally, we applied the proposed algorithm into two scientific workflow applications that have different complex structures and made a simulation. By setting different restrict degree of both deadline and budget, we observed and analyzed the performance of scheduling algorithm. We also analyzed the influence to the performance of the algorithm based on different iterations and different initial population scale.