The scale of cluster has gradually expanded and the performance has be increasing.As a key technology,cluster scheduling has become the focus and difficulty of the research.Traditional scheduling algorithm is not only difficult to solve the problem of performance mismatch between job requirements and cluster nodes,but also easily lead to uneven load among the cluster nodes,which greatly limits the performance of cluster system.Therefore,the intelligent scheduling algorithm based on the simulation of natural ecology has gradually been favored by the people,such as the typical genetic algorithm and ant colony algorithm.Ant colony algorithm,pheromone update is very important,it has two parts,one part is the path of the original pheromone residues,and the other part of the ants after the release of the pheromone.The thesis puts forward the IACO algorithm,and the algorithm improves the two aspects of the algorithm,by introducing performance matching factor and load balancing factor to improve ant the path through the release of the pheromone to solve job requirements and the cluster nodes resource performance does not match the problem,and the residual pheromone on the original path to solve the problem of load imbalance among the cluster nodes Firstly,analyze,test and compare the traditional scheduler algorithm and a part of intelligence algorithm,prove that Shuffled Frog Leaping Algorithm have a good performance.The experimental results show that the IACO algorithm reduces the execution time,the CPU utilization rate is improved,and the load of each node is more balanced.But the execution time and CPU utilization is not particularly desirable.The main reason is ant colony algorithm early information element deficient,the convergence speed is slow.The genetic algorithm has fast global searching ability when it is searching for a widerange of solutions.So the thesis puts forward the GA-IACO algorithm and the improved ant colony algorithm and genetic algorithm together to solve the cluster scheduling problem.In the early stage,genetic algorithm is adopted to make full use of the advantages of genetic algorithm to generate the initial solution,and to stay in a better path.In the later stage of the algorithm,the improved ant colony algorithm is used to make full use of the positive feedback of the ant colony algorithm to find the optimal solution of resource allocation.The experimental results show that the GA-IACO algorithm has a significant improvement in the efficiency of the implementation,and achieved satisfactory results.In order to test the improved algorithm better,the IACO-SS scheduling software is designed and installed in the cluster.Experiments show that the improved algorithm reduces the cluster scheduling execution time,improve the utilization rate of CPU,so that each cluster node load more balanced. |