Abstract: To solve the service dynamic selection problem in cloud computing environment,a fitness function which considers both the response time and the cost was designed,and an estimation of distribution-shuffled frog leaping algorithm was proposed to solve the problem of service dynamic selection.On the basis of leapfrog algorithm,evolutionary operators of leapfrog algorithm was redefined by drawing crossover operation of genetic algorithm,and distribution estimation evolutionary strategy was introduced to improve frog update mode of the leapfrog algorithms,so that the new improved algorithm has a more comprehensive learning ability and it can effectively avoid the local optimum.Simulation results demonstrate the feasibility and effectiveness of the proposed algorithm,and compared with the leapfrog algorithm and estimation of distribution algorithms,the convergence performance and optimization capabilities of the proposed algorithm are improved,and it can better solve the service dynamic selection problem in cloud computing environment.

Key words: Cloud computing,Service dynamic selection,QoS,Evolutionary operators,Fitness function,Probabilistic model