Research On DAG Task Scheduling Dynamic Fusion Based On Genetic Ant Colony Algorithm In Cloud Computing Environment

With the development of the computer science, technology of the computer and the network of computer, a new calculation method-cloud computing gradually comes into the sight of person. Because cloud computing has strong expansible and practicability ability, the scholar up it to the top of the cloud computing in academia and industry, so that cloud computing has got an unprecedented development. The cloud computing model is a kind of producer consumer model that is resource delivery and usage model, it is a great convenience to consumers, but also to expand the interests of producers, so the cloud computing model is bound to be the trend of the future. However, a large number of the data and the calculation of the producers are concentrated in the cloud, then it will consume more and more resources and energy, so that it will bring the problem of unstable operation and the waste of energy and system,which is the main reason of task scheduling that is not reasonable. Therefore, the best way to achieve cloud computing system performance and optimization of energy consumption is to study the design of efficient task scheduling algorithm.The main task of the current scheduling algorithm concluding:based on the cost of the task scheduling algorithm, based on the Petri network of the task scheduling algorithm,based on the Agent of the task of scheduling algorithm, a heuristic algorithm of the task scheduling algorithm and other task scheduling algorithm. This paper will mainly study the heuristic task scheduling algorithm, then we will use this to beg the optimal solution of the cloud computing task scheduling problem. The forming process of genetic algorithm is that the researchers breed the process of the evolution and mimic of the biological animals, the genetic algorithm with the advantages of the diversity of the population, the strong global searching ability and parallelism and adaptive, but the genetic algorithm has shortcomings to solve the best solutions of the task scheduling problem; however,the formation process of ant colony algorithm is that the scholars simulation of ant foraging process,the ant colony algorithm has the same advantages of distributed parallel, adaptive, and it can provide the feedback information, also it has strong robustness and performance advantages, but the ant colony algorithm searches results of the task scheduling problem of a long time, also it is easy to fall into local optimal solution, which is a big problem of the ant colony algorithm.On the basis of this, we are necessary to improve the genetic algorithm, on the one hand, we can make a very good response to the change of work flow tasks, if task changes then we will update the DAG map; on the other hand, we will add the user satisfaction in the fitness function of genetic algorithm, as a constraint condition of the genetic algorithm,then it will improve satisfaction degree;finally, we will consider the clustering process in the resource of task scheduling, to achieve the maximization of resource utilization. Also, we made the necessary improvements of the ant colony algorithm, we will updating strategy and local update strategy based on the use of improved algorithm, then it will be better to the query performance optimization scheme, at the same time we made the necessary improvements of the volatile coefficient of the ant colony algorithm, so that it can be dynamically adjusted with the development of the ant colony algorithm iterative process in the volatile coefficient, then the simple scaling of the resource node information, the load of the individual resource is too large, so that the cloud computing task scheduling can achieve the balancing of load goals, and it can accelerate the convergence speed of ant colony algorithm.Although we made the necessary improvements of genetic algorithm and ant colony algorithm,but it is influenced by their own algorithm, the efficiency of the algorithm is still not formed quantitative change. Therefore,based on the heuristic task scheduling algorithm, based on the work flow task model, we realize the integration of the dynamic the improvement of genetic algorithm and ant colony algorithm, we will change the genetic algorithm into ant colony algorithm in the best time, this paper first to the area by GA,then we will choose the optimal individual top 10% that is searching by genetic algorithm,then we use the ant colony algorithm, as the implementation of the initial pheromone of ant colony algorithm, and then through the implementation of exact solutions, cloud computing task scheduling. Proved by experiments, so as to realize the exact solution well,and we need less time to search for the optimal solution.The main idea is that we will dynamic integration of our development of the ant colony algorithm and the genetic algorithm, to avoid the limitations and shortcomings of their respective algorithm itself, we only use the advantages of their own algorithm each, to form a more perfect heuristic intelligent optimization algorithm.Finally, this paper used cloud computing data center task scheduling simulation system of CloudSim,according to the improved genetic algorithm, the improved ant colony algorithm, and the ant colony algorithm integration of the simulation comparison, simulation results show that,the improved algorithm is better than genetic algorithm and ant colony algorithm only.