Analysis And Improvement Of Task Scheduling Algorithm Based On Cloud Computing Environment

As a new business computing model,cloud computing has been widely concerned by many large-scale IT companies,and develops rapidly under research and promotion of many application industries.Cloud computing uses virtualization technology to convert large-scale and heterogeneous hardware infrastructure into unified-management virtual resource pool,and users can easily get various types of flexible use of virtual resources without concerning for their specific location and the underlying configuration.Due to the huge user group of cloud computing system and dealing with the massive task and data,how to properly allocate virtual resources is significant to the research of cloud computing scheduling algorithm to reduce processing time and expense of users' tasks and achieve load balancing.Contrary to slow convergence,poor global search and easily falling into stasis of the BACO,this paper analyzes it detailedly and optimizes the search process with chaotic theory,and proposes an improved ant colony task scheduling algorithm.The algorithm uses chaotic sequence to increase the randomness of the transition probability,and it adjust pheromone's concentration on the optimum path and other paths.In addition,the algorithm also takes into account the different situations in the searching process and modifies the pheromone updating rule,which avoids the search falling into stasis and improves the system's load balancing.In the full consideration of the quality of service needs of different users' tasks,this paper proposes the multi-QoS constraints task scheduling algorithm based on ACO model with the advantages of ant colony optimization.The algorithm defines the mathematical models of time,cost,reliability,availability and security and builds fitness function by setting different weights.In the designing of the heuristic function,this algorithm considers the computational performance of resource nodes,and it improves calculation cost and the system's load balancing.In addition,it uses genetic crossover operator and designes different updating rules of local and global updating process,which improves the global searching ability of the algorithm to a large extent.Finally,the paper use CloudSim platform to simulate cloud computing environment,and it tests the two algorithms metioned and analyzes experimental results.From the first experimental result we can know that the improved ant colony optimization can shorten the average completion time and reduce execution cost of the tasks,and it can also make thesystem's load more balancing.The second experimental result show that the multi-QoS constraints task scheduling algorithm based on ACO model has improved in terms of computational cost and the system's load,while providing a good reference for relevant research.