The cloud computing resource pool is composed of a large number o f resource nodes with different performance.However,as the number of users gradually increases,the demand increases,and how to effi ciently allocate large-scale cloud tasks to a li mited number of resource nodes,And to achieve load balancing,is an i mportant issue that cloud computing needs to study.Task scheduling is a typical NP-compl ete problem,which makes the group intelligence algorithm has become a research hotspot for many scholars to explore and solve cloud computing task scheduling strategy.Particle swar m optimization(PSO)is one of the mainstream intelligent algorithms to solve the task scheduling opti mization probl em of workflow system in cloud computing environment.Because the particle swarm opti mization algorithm i s easy to fall into local optimum,leading to scheduling program execution time and high cost defect s,so this dissertation opti mizes the particle swarm task scheduling algorithm based on the traditional adaptive inertia weight,and accurately describes the particle state to i mprove The self-adaptive inertial weights are proposed,and a task scheduling algorithm is proposed that i mproves t he calculation of the success of individual particles.The i mproved adaptive inertia weight can more accurately adjust the particle velocity,effectively i mprove the defect s of the traditional particle swarm algorithm that can easily fall into the local opti mal solution,and can get t he schedul ing scheme that the execution time and the cost both satisfy the expectation.Experi mental results show that t he i mproved algorithm has less fluctuations in the convergence speed curve and the convergence accuracy is i mproved significantly.It is an effective t ask scheduling algorithm.Finally,by analyzing the characteristics of the i mproved PSO and ant colony algorithm,we propose a task scheduling algorithm(ITAIWPSACO)that integrates the two advant ages.Experimental results show that the fusion algorithm further i mproves the overall efficiency of cloud computing platfor m task scheduling strategy. |