LAUSR

Abstract Blockchain cloud manufacturing is an emerging service-oriented paradigm that takes many advantages to cloud manufacturing. Previous researchers have studied from eradicating third party trust problem among service providers and customers by decentralized data record and involving smart contract. However, the existing paradigm still suffers from third party trust problem among service-related participants and cloud manufacturing provider since service is solely managed by cloud manufacturing provider without supervision and service related participants cannot know whether composed service provision which indeed is NP-hard problem is optimal. In this paper, the architecture of blockchain cloud manufacturing system is proposed with adoption of several cloud manufacturing providers and vote mechanism to vote for optimal composed service provision on candidate composed service scheme firstly calculated by one cloud manufacturing provider. According to customer satisfaction-based service match and quality control, a system framework of composed service performance evaluation is desifned to incorporate the vote mechanism. Firstly, customer satisfaction dimensions as evaluation dimensions are extracted from customer text reviews based on latent dirichlet allocation (LDA) and sub-dimensions under each customer satisfaction dimension are set up based on professional field knowledge. Secondly, overall customer satisfaction rating of reviews and sentiment of reviews towards customer satisfaction dimensions are analysed by three score scheme, bag of words (BOW) and support vector machine (SVM) classification. Thirdly, the sentiment effect and category of customer satisfaction dimension on overall customer satisfaction are identified with the application of neural network and Kano model. Fourthly, in consideration of category characteristic, overall performance of customer satisfaction dimension is measured as distance and using proposed Logit-SVM method. Service selection standard according to value of customer satisfaction is set up. Finally, one empirical study is presented to verify the feasibility of proposed system framework, and proposed model is further discussed.