Research On Task Scheduling In Cloud Computing Based On Fusing Algorithm Of Genetic And Ant Colony

Cloud computing is a new computer network technology and service model, it provides efficient and convenient services with its high scalability and ease of use for people’s life and scientific research, such as large data and high-performance computing. The basic principle of cloud computing is divided the task of the program submitted by the user into several sub-tasks firstly, and then transfer to a large computing centers which composition by multiple server. After distribution, scheduling, calculation and analysis, the consolidated results will return back to the user. However, when the computer in cloud dealing with large amounts of data, due to the factors such as inability to predict the arrival time of the task, different efficiency and response time between different machines in dealing with different task, frequent communications between the calculation nodes and so on,the power consumption of data center and high-performance computing platforms becomes increasingly larger. Explained that the main cause of energy consumption unreasonable task scheduling. Therefore, it is meaningful to design an energy-efficient scheduling method that can efficient implementation of cloud computing task scheduling and reduce energy consumption which caused by each computing node.On the basis of analysis algorithms and computational models of task scheduling in cloud environment, existing heuristic algorithm was improved in two ways: 1. For the defects of GA’s late slow convergence, ACA’s slow solving speed in early and easy fall into local optimum, proposed an improved algorithm which dynamics fusing of genetic algorithm and ant colony algorithm, abbreviated as D-GAACA. 2. Designed an energy-efficient scheduling model and put the double constraint factor of time-energy as its fitness function. The model can be optimized simultaneously from total time to complete and compute nodes energy consumption(calculate energy, communication energy, transport energy consumption) of the tasks. Finally, achieve new scheduling algorithms and models in the Cloudsim3.0.3 cloud simulation platform. For Ensure the efficiency of the algorithm, experimental analyzes the optimal combination of relevant parameters of GA in cloud environment firstly, and then compare the D-GAACA with others scheduling algorithm, such as ant colony optimization algorithm(ACO), FCFS algorithm, and round robin algorithm(RR). The results proved that the new algorithm can solve the energy-efficient scheduling problems better.