Optimization Of Cloud Manufacturing Service Composition Based On Improved Sparrow Algorithm

Cloud manufacturing mode is a user-centered,highly integrated and intelligent manufacturing model.In order to meet the rapid development of the manufacturing industry and the dynamic changes in user business needs,the service composition in cloud manufacturing can combine single functional cloud services into complex value-added cloud services.High-quality cloud manufacturing service composition can enhance user experience and enterprise competitiveness.However,when the cloud manufacturing system provides more similar manufacturing cloud services with different nonfunctional attributes,it is often faced with a large number of candidate combinations and difficult to find the optimal solution within the feasible time.Therefore,how to search quickly from a large number of service composition schemes and obtain the service composition most in line with users’ expectations is a challenging task.The sparrow search algorithm proposed in recent years has the characteristics of better convergence speed and optimization effect,which is suitable for solving the cloud manufacturing service composition optimization problem with huge search space.Based on the background of cloud manufacturing,this paper evaluates the quality of composite services by considering the important indicators of the service composition optimization process,and propose different versions of sparrow search algorithm for different application scenarios.The main research contents are as follows:(1)Firstly,this paper analyzes the characteristics of the cloud manufacturing service composition optimization problem,selects four key optimization indicators from the perspective of resource providers and cloud platforms,including manufacturing time,cost,reliability and transaction quality.Based on this,an evaluation model based on service quality is constructed,and a penalty function is introduced to reduce the impact of strict constraints on the final optimization effect;At the same time,a dual objective service composition evaluation model based on service quality and energy consumption is established by analyzing energy consumption composition.(2)In the early stage of solving the service quality evaluation model,K-means clustering is used to select the candidate services with better quality.In the later stage of optimization,when Sparrow search algorithm deals with the single objective cloud manufacturing service composition problem,it has the disadvantages of loss of group diversity and lack of global search ability due to its local search mechanism,which reduces the quality of service composition solutions.Therefore,chaotic mapping is used in the initial stage to ensure the generation of population diversity.Then,t distribution and differential perturbation are used for position updating in the late evolution period to boost the global optimization ability of the algorithm.Finally,the entire process achieves a balance between global search and local exploration by adjusting the number of discoverers and followers in the group through adaptive factors.The performance of this algorithm is evaluated using the benchmark functions,and the results confirm the effectiveness of the improved strategy proposed in this paper.Moreover,the experiments on the service composition problem further demonstrate that the algorithm proposed in this paper has good search performance and optimization accuracy when solving the service quality evaluation model.(3)Finally,considering the uneven distribution and lack of diversity of non dominated solutions in maintaining external archives and selecting the optimal individual to guide population updates when dealing with a dual objective evaluation model based on service quality and energy consumption,the sparrow search algorithm adopts a grid method to maintain the distribution of the solution sets from the perspective of archive updates,and improves the diversity of solutions based on a roulette wheel strategy,expanding it into a multi-objective sparrow algorithm.At the same time,an enhanced global search strategy and polynomial mutation operator are designed to address the problem of search stagnation and insufficient search range for later position updates among the discoverers in their own population.Through the benchmark function and cloud manufacturing service composition problem,the performance of the improved algorithm is identified and analyzed.The experiment indicates that the proposed algorithm has better convergence distribution and diversity,and enhances the quality and convergence stability of the solution to the double objective service composition optimization problem.