Challenges in Developing a Computational Platform to Integrate Data Analytics With Simulation-Based Optimization

The focus of the work presented in this paper is to identify and find possible solutions for major implementation challenges in designing a computational platform for integrating data analytics paradigm with the simulation-based optimization technique to facilitate the modeling of a smart manufacturing system. A simulation model of a manufacturing system generates real-time monitoring data for machine status and these data are then mined by data mining algorithms to discover hidden knowledge that might not be predefined in the simulation model. The new knowledge is then fed into the simulation model such that the model adapts and evolves, and eventually it can predict future status. This procedure involves heterogeneous modeling techniques, information exchange among different tools, as well as model composition and interaction. We extend an early presented “Hypercube” information model that was specifically developed for the purpose of formal representation of smart manufacturing systems, in order to harmonize the information required by the simulation modeling tool and the data analytics tool. A strong emphasis is given to emerging areas of multi-domain and multiscale modeling by means of integration and interoperability between existing modeling tools and technologies. A specific case study related to preventive and predictive maintenance of a typical manufacturing system has been elaborated in the paper as the initial scope and application area in order to illustrate and validate the proposed computational framework.