Research on recommendation and interaction strategies based on resource similarity in the manufacturing ecosystem. (October 2020)

• Record Type: Journal Article
• Authors: Li, Jiming; Zhang, Yingfeng; Qian, Cheng; Ma, Shuaiyin; Zhang, Geng
• Is Part Of: Advanced engineering informatics. Volume 46(2020)
• Publication Date: 2020-10
• Subjects: Recommendation system -- Resource interaction -- Manufacturing ecosystem -- Resource similarity; Computer-aided engineering -- Periodicals; Engineering -- Data processing -- Periodicals; 620.00285
• DOI: 10.1016/j.aei.2020.101183
• Languages: English
• ISSNs: 1474-0346

Abstract

Abstract: Nowadays, the thriving of the manufacturing ecosystems (ME) driven by the increasing competition in industrial markets, the ubiquitous implementation of intelligent systems, and the more frequent collaboration among manufacturing enterprises. During the practice of the system upgrade, it is increasingly noted that the redundancy of manufacturing resources and the inefficiency in resource configuration are the major obstacles to achieving satisfying value-creation within ME, which also result in cumbersome decision making (DM) in the problems of requirement-service configuration (RSC) and collaborative production. To address these issues, the research on resource recommendation and interaction is carried out. Firstly, the resource similarity models for autonomous resource filtering brace the whole DM mechanism in RSC and push the most suitable resource to the host automatically. Then, the interaction model provides a self-organized production mode without human intervention. The blindness, lag, and unfairness in the manual communication is eliminated by the Machine to Machine (M2M) interaction and automatic coordination. Besides, an NLP-based machine learning algorithm is introduced for quantifying semantic distance and measuring the differences between orders. Composed by these models, a total solution named Industry-Chat (I-Chat) emerges. With the help of that, production resources can be scheduled and managed autonomously and the order-based production processes could be promoted seamlessly. Thus, an improved industrial ecosystem with automatic DM and self-organization for future intelligent manufacturing is realized. The practicability of the research is verified by a case study. The results show that the production cost is reduced by 12%, the resource utilization rate is improved and its economic value is demonstrated.