Research On Manufacturing Service Composition Optimization Models And Algorithms In Cloud Manufacturing

Cloud manufacturing is a new model of networked manufacturing,which is service-oriented.With the emergence of cloud computing technology,many services with the same functionalities and different non-functionalities occur in cloud manufacturing system.Due to the growing complexity of customer requirements and the increasing scale of manufacturing services,how to select and combine the appropriate services to compose manufacturing service composition solution to meet the complex demand of the customer has become a growing concern.This study aims to solve the problem of manufacturing service composition optimization in cloud manufacturing system,which is based on quality of service(QoS).To achieve this goal,we propose a new three-dimensional manufacturing service composition model under scheduling mode,which can solve the problems of service aggregation,service selection,and service scheduling simultaneously.Futher,to extend the above model,we propose a correlation-aware manufacturing service composition model under crowdsourcing mode,which considers crowdsourcing and correlations among manufacturing services.Finally,different improved Flower Pollination Algorithms(FPA)are employed to solve the corresponding model and their performance are proved by experiments.Thus,the main contributions of this paper are concluded as follows:1.We propose a new manufacturing service composition model based on three-dimensional structure under scheduling mode.First,the proposed model not only presents different methods for calculating the transportation time and transportation cost under various structures but also solves the three-dimensional composition optimization problem,including service aggregation,service selection,and service scheduling simultaneously.Second,an improved FPA is proposed to solve the three-dimensional composition optimization problem using a matrix-based representation scheme.The mutation operator and crossover operator of the Differential Evolution Algorithm are also used to extend the basic FPA to improve its performance.Finally,compared to Genetic Algorithm,Differential Evolution Algorithm,and basic FPA,the experimental results confirm that the proposed method demonstrates superior performance than other meta-heuristic algorithms and can obtain better manufacturing service composition solutions.2.We propose a new manufacturing service composition model based on service correlation under the crowdsourcing mode.First,the proposed model considers both crowdsourcing and the influence that service correlation exerts on the QoS value of manufacturing service.Second,to address the problem of multi-objective optimization,we employ an extended FPA to obtain the optimal service composition solution,where it not only utilizes the adaptive parameters but also integrates with Genetic Algorithm.Finally,a case study was conducted to illustrate the practicality and effectiveness of the proposed method compared with Genetic Algorithm,Differential Evolution Algorithm,and basic FPA.