Research On Service Personalized Recommendation And Composition In Cloud Manufacturing Environment

With the composition of information technology and advanced manufacturing mode,cloud manufacturing has been presented to promote the manufacturing transformation from manufacturing-based mode to service-based mode.In cloud manufacturing environment,enterprise can effectively obtain the valuable manufacturing service,due to tightly connect distributed manufacturing resources and effectively sharing.However,the openness and sociability of cloud manufacturing will bring some challenges for the manufacturing service recommendation and composition.First,state-of-the-art manufacturing service recommendation methods have ignored the dynamic and global provider reputation value and user trust value,making the service recommendation inaccurate and ineffective.Second,the fuzzy and complexity of cloud manufacturing make the manufacturing service composition more difficult that is lack of an effective mathematical model to accurately quantify fuzzy and complex manufacturing service composition problem.To overcome the above drawbacks,we set the goal of accomplishing manufacturing service sharing,personalized recommendation,and service composition in cloud manufacturing environment.A new reputation evaluation method is proposed to obtain global and dynamic reputation values of providers and trust values of users by combining time-aware and HITS algorithm.In order to recommend suitable manufacturing service,we present a hybrid recommendation method that collaborative filtering technology is combined with reputation values of provider,trust value of users,and enterprise features.In addition,we employ the triangular fuzzy number to build a new mathematical model for fuzzy and complex manufacturing service composition problem.Furthermore,we modify the standard Biogeography-based optimization to solve the proposed model and make the performance of standard Biogeography-based optimization better.The main contribution of this paper is concluded as follows:1.We propose a new reputation evaluation method called time-aware HITS algorithm in cloud manufacturing environment.According to openness and dynamic of cloud manufacturing,we take full advantage of HITS algorithm to effectively and efficiently calculate the global reputation values of providers and trust values of users.Furthermore,the temporal dynamics are taken into account for computing the global and dynamic reputation value,making the result more reasonable and accurate.2.We propose a hybrid recommendation method for manufacturing service personalized recommendation in cloud manufacturing environment.In addition to the reputation value,the enterprise features are further taken into account to meet the personalized requirements of enterprises and make the service recommendation more accurate.3.We propose a new fuzzy quality of service(QoS)-aware multi-objective mathematical model for evaluating the global QoS value of manufacturing service composition in cloud manufacturing.In addition,the BBO algorithm is extended to effectively obtain the solution of manufacturing service composition with an optimal fuzzy QoS value,by improving its standard migration and mutation operators and introducing a new operator called the invasion operator.Finally,two set of comparison experiments are carried out to illustrate the practicality and effectiveness of our proposed manufacturing service recommendation method and service composition method,respectively.