Research On Swarm Cluster Scheduling Strategy Based On Improved Shuffled Frog Leaping Algorithm

In recent years,with the development of cloud computing technology,distributed computing technology has become the first choice of back-end technology solutions for major enterprises in today's era.However,with the popularization and application of distributed technology,the problem of resource load balancing and improving resource utilization in cluster has gradually become one of the development bottlenecks.In recent years,container virtualization technology has improved the low utilization of resources in virtual machine technology,but the problem of load balancing still exists.Based on the current popular Docker container technology,this paper studies the shortcomings of the scheduling strategy in Docker Swarm cluster tools and the unbalanced load of nodes.Firstly,the shortcomings of the existing scheduling strategies are analyzed,and then a dual strategy hybrid leapfrog algorithm based on threshold is proposed.Finally,it is used to solve the problems of load imbalance,low scheduling efficiency and low resource utilization caused by unreasonable container scheduling.The main work includes the following aspects:First of all,the current mainstream docker cluster management tools are studied and compared,and the reasons for choosing docker swarm cluster tools as the experimental platform are explained.At the same time,the shortcomings of the built-in scheduling strategy of docker swarm are analyzed,and a hybrid leapfrog algorithm is introduced for improvement.However,the standard hybrid leapfrog algorithm itself has the problems of slow convergence speed and easy to fall into the local optimal solution.Therefore,a dual strategy local search strategy based on the threshold is proposed.It is also an individual cross mutation strategy to determine the group mutual learning strategy by determining the individual dispersion in the population.Experiments with a series of standard test functions show that the improved algorithm can effectively improve the convergence speed and local search ability of the standard hybrid leapfrog algorithm.Secondly,through the mathematical model of container scheduling problem,combined with the characteristics of docker container scheduling strategy,a dual strategy hybrid leapfrog algorithm docker based on threshold is designed Swarm scheduling strategy,which evaluates the load of nodes from five aspects of CPU,memory,hard disk space,IO load,network load and analyzes the resource requirements of the container on demand,so that the container adopts batch execution scheduling to improve the efficiency of scheduling.Finally,through the comparison of multiple experimental tests,it is found that the improved hybrid leapfrog algorithm is more efficient than the standard hybrid leapfrog algorithm,and the docker swarm container scheduling strategy based on the threshold hybrid leapfrog algorithm is more balanced and efficient than the spread scheduling strategy built in docker swarm.