Abstract
In recent years, the manufacturing industry is changing its way of production enabled by emerging technologies. Enterprises publish visualised resources as services on the cloud. To fulfil a complex manufacturing request, services can be composed together. However, how to compose proper services among a large number of services becomes a challenge. This paper presents a service composition graph model to deal with the challenge by using a graph database. Specifically, the service composition graph model is stored in the Neo4j graph database. Services and their inputs/outputs are stored as nodes in a directed bipartite graph that are connected with edges. The weights of edges are synthesised via a fuzzy control system according to the QoS of services. Possible compositions of services are calculated and pre-composed. When the manufacturing task arrives, an extended Dijkstra algorithm is used to find a solution. Experimental results show that the approach could return a solution satisfying both functional and QoS requirements within a short time.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
