Research On Cloud Service Composition Optimization Based On Improved Moth-flame Optimization Algorithm

With the rapid development of cloud computing,a single service can no longer meet the needs of realistic functions.A group of atomic services needs to be selected in the resource pool to cooperate with each other.The concept of cloud service composition has thus come.The cloud service composition problem is regarded as an NP-hard problem,and it has also become a highly-researched technical difficulty in the cloud computing system.The moth-flame optimization algorithm(MFO)has been proposed in recent years and has the advantages of easy operation,simple parameters,and easy programming.Some scholars have tried to use it to solve the problem of cloud service composition,and have achieved some results.However,this algorithm has the disadvantage of fast convergence of the moth-flame optimization algorithm,the lack of intra-population communication and the prone to local extremities.As the complexity of the scene increases,the reliability,availability,and fitness values in the experiment decrease rapidly,this algorithm can't meet the needs of users well.Therefore,this thesis attempts to propose an improved moth-flame optimization algorithm to in-depth research the problem of QoS service composition in the cloud computing.Firstly,the research background and significance are introduced in this thesis,and the research status of intelligent heuristic algorithms in the QoS cloud service composition chronologically is elaborated.After that,the basic knowledge of cloud computing and service composition is briefly introduced,and then the common intelligent heuristic algorithms of cloud service composition are explained,focusing on genetic algorithm,particle swarm algorithm and cuckoo search algorithm.Against the shortcomings of the original algorithm,an improved logarithmic spiral function moth-flame optimization algorithm(ISMFO)is proposed in this thesis.This algorithm uses the adaptive weight method to improve the flight mechanism of the logarithmic spiral function.When moths approach the flames,the adaptive weight will decrease in order,thereby the local optimization ability of the moth is improved and falling into the local optimal solution is avoided,then this algorithm restrict the moth's search space to prevent flying to invalid space and affect the optimization efficiency.After that,in order to expand the optimization range of moths and enhance the internal communication of the population,this thesis proposes cuckoo search and moth-flame joint optimization algorithm(CSMFO).This algorithm adds a moth random migration mechanism after the moth position is updated,which increases the randomness of the moth position update,and enables the moth to perform a better global search.Combining the above two improve algorithms,this thesis proposes new moth-flame optimization algorithm(NMFO).In this thesis,ten test functions are used for testing on matlab platform.The proposed ISMFO,CSMFO and NMFO are compared with moth-flame optimization algorithm(MFO),cuckoo search algorithm(CS)and particle swarm algorithm(PSO).The test results show that ISMFO,CSMFO and NMFO optimization effects are better than MFO,CS and PSO.The NMFO optimization effect is far superior to the other five algorithms,so NMFO is used for cloud service composition simulation.Finally,three cloud service combination scenarios are used for simulation on the cloudSim platform,then the execution time,service cost,reliability,availability in the QoS attributes and fitness value are used for performance evaluation.The new moth-flame optimization algorithm(NMFO)is compared with the moth-flame optimization algorithm(MFO),cuckoo search algorithm(CS)and particle swarm algorithm(PSO)in QoS cloud service composition simulation experiment.Experimental data shows that the reliability,availability,and fitness values of NMFO in three scenarios are higher than those of the other three algorithms,which shows that new moth-flame optimization algorithm(NMFO)is used in cloud service composition can improve the quality of cloud service composition and user satisfaction.