Grey Wolf Optimization Algorithm Based On Opposition-Based Learning And Trial Perception Strategy

The problem of cloud task scheduling in the field of cloud computing has been a hot topic of recent research by scholars.The essence is to find a better algorithm to assign the tasks submitted by the user to the cloud data platform to the current resources for execution,with the ultimate goal of efficient utilization of resources and load balancing in the cloud platform.The swarm intelligence optimization algorithm in task scheduling is inspired by the various behaviors of animals in nature,which have been modelled.Compared with traditional optimization algorithms,swarm intelligence optimization algorithms have higher performance in solving complex optimization problems.Grey Wolf optimization algorithm(GWO)is a new swarm intelligence optimization algorithm proposed in 2014.Due to the advantages of fewer parameters and fast convergence,GWO has been widely used in function optimization,UAV path planning,and other fields.However,grey Wolf optimization algorithms suffer from some problems,such as poor population diversity and the tendency to get trapped in local optimal solutions.In this paper,we propose a modified Grey Wolf optimization algorithm that includes the following three aspects:First,the algorithm is optimized at the initial stage of the Grey Wolf optimization algorithm,and the dependence of the algorithm on the initial population is eliminated by adversarial learning.The classical Grey Wolf optimization algorithm randomly generates when initializing the search agent,which means that the algorithm strongly depends on the quality of the initial population.If the initialized agents are all distributed near the local optimal solution and far away from the global optimal solution,it is easy to fall into the local optimal solution in the process of successive iterations of the algorithm and it is difficult to search for the global optimal solution in the later iteration of the algorithm.The adversarial-based learning initialization population can improve the quality of the initial search agent at the beginning of the algorithm by generating inverse search agents in the case of randomly generated search agents and eliminating search agents with poor fitness by comparing fitness functions.Second,it fine-tunes the problem where the GWO algorithm gets stuck in a local optimal solution.In the middle of the algorithm,with the increase of the number of iterations,if the leadership in the search agent drove into the direction of the local optimal solution,the group in the algorithm would follow the leadership to drive into the direction of the local optimal solution,resulting in the algorithm falling into the local optimum and missing the objective existence of the global optimal solution.In the process of iteration,the grey Wolf optimization algorithm based on the trial perception search strategy generates the trial factor around the search agent and participates in the calculation of the fitness and the election of the leadership,so as to enhance the exploration ability of the leadership and drive towards the direction of the global optimal solution.Finally,the optimized Grey Wolf optimization algorithm was tested on the matlab simulation platform using the standard test function library Virtual Library of Simulation Experiments,and two metrics were mainly selected for testing..The first one is the function convergence accuracy.Here,the mean,standard deviation,maximum and minimum are used as the main reference parameters,the second is the convergence rate of the function,and the evaluation metric is the line graph of the function over the specified number of iterations..In both test metrics,the modified Grey Wolf optimization algorithm achieves better results.