Modeling and Analyzing Performance of a Production Unit Using Dynamic Bayesian Networks

The aim of this paper is to formulate a quantitative integrated model of how quality and productivity performances of a production system are interrelated. Indeed, productivity and quality, some of most important objectives of a production system have been studied separately since decades whereas studies are demonstrating a close interaction between them nowadays. Such an integrated model will be beneficial to engineers during design and/or operation stages of the system because it can be used to set up or to assess overall performance measures such as: total production rate, effective production rate, machines availability, inspection policies performance, etc. Dynamic Bayesian network will be used as the underlying mathematical tool to describe the dynamics of the state of the system as they are well suited for the representation of stochastic processes (machine failures, quality failures, etc.).