Research On The Task Scheduling Strategy Based On Multi-Objective Optimization In Cloud Environment

Task scheduling strategy is an inevitable problem faced by cloud computing,for which good system operation efficiency and high user satisfaction require a better task scheduling strategy.In this case,how to optimize the task scheduling strategy in cloud computing so as to improve the system operation efficiency and user satisfaction has always been a research hotspot and difficulty.At present,there are many types of task scheduling strategies in cloud computing,but most of them only have one optimization target,i.e.one of load balancing,task completion time,cost or energy consumption.Multi-target optimization scheduling strategies,such as algorithm with task completion time,cost and load balancing as the common optimization target,are still relatively few and lack of depth and comprehensiveness.In addition,the ant colony algorithm is widely used in solving NP-hard problems,while cloud computing task scheduling is included.However,the standard ant colony algorithm has some problems,such as easily falling into local optimum and slow convergence in the absence of initial pheromone,and the improved one only has a single optimization target when applied to cloud computing task scheduling.Moreover,the ant colony algorithms with task completion time,cost and load balancing as common optimization targets are still relatively few.After researching and analyzing the ant colony algorithm and the existing task scheduling model,this thesis proposes a task scheduling strategy aiming at optimizing load balancing,cost and task completion time based on the ant colony algorithm.First,the three targets are separately modeled,and by use of linear weighting,the objective function is constructed after normalization to convert the three targets into a single target.After that,dynamic adaptation of pheromone Q is introduced,and the pheromone update strategy of ant colony algorithm is improved by combining the constructed objective function,and then the task waiting time factor is incorporated into the expected heuristic function,so that the ant algorithm can avoid falling into local optimum and its load balancing metrics can be improved.After the algorithm is proposed,this thesis verifies it on the CloudSim simulation platform.The experimental results show that the multi-target optimization task scheduling strategy based on ant colony algorithm has a certain improvement on load balancing,cost and task completion time.