Research On Composition Optimization Of Cloud Manufacturing Service Based On Improved Butterfly Optimization Algorithm

With China's industrialization entering the late stage,the traditional manufacturing mode is gradually changing to the intelligent manufacturing mode.Under the background of the new industrial era,artificial intelligence,Internet technology and manufacturing environment are integrated to produce a new cloud manufacturing mode.Cloud manufacturing service composition combines a large number of decentralized and heterogeneous manufacturing services into complete manufacturing tasks according to user requirements.How to select and match the most suitable cloud manufacturing service composition for user needs has become a key problem in the field of cloud manufacturing.Cloud manufacturing service composition is usually a discontinuous and non derivative NP hard combinatorial optimization problem,which is suitable for solving by swarm intelligent optimization algorithm.Butterfly optimization algorithm is a new swarm intelligence optimization algorithm proposed in recent years.Aiming at the disadvantage that the butterfly optimization algorithm is easy to fall into the optimum,this paper introduces the good point set and the opposition-based learning strategy to improve the butterfly optimization algorithm.The good point set makes the initialization of the population uniform,and the opposition-based learning increases the randomness of the population updating,so as to improve the convergence of the butterfly optimization algorithm.Experiments show that the improved algorithm has good convergence and stability.Integrating the reliability,availability and throughput of cloud manufacturing service composition,an optimization model of cloud manufacturing service composition based on QoS is proposed.The model solves the problem of unreasonable QoS attribute of cloud manufacturing service composition,and can provide users with a higher level of manufacturing services.The butterfly optimization algorithm based on good point set and reverse learning is used to solve the model.The experimental results show that the algorithm is effective.A multi-objective butterfly optimization algorithm is proposed by introducing external elite archives and grid method.The grid method can obtain the uniform solution set and the external elite archives can promote the convergence of the solution.Compared with the classical multi-objective algorithm on several benchmark problems,the excellent performance of the multi-objective butterfly optimization algorithm is verified.Considering the time,cost and manufacturing level of cloud manufacturing service composition,a multi-objective optimization model of cloud manufacturing service composition based on comprehensive manufacturing level is proposed.The model solves the problem of excessive total cost of user manufacturing and can realize more effective service allocation.Using the above multi-objective butterfly optimization algorithm to solve the model,the experimental results show that it can obtain a better solution than the comparison algorithms.