In order to achieve responsive manufacturing control systems greater understanding of the manufacturing system control problem needs to be gained. More specifically, techniques need to be developed that can be used to model and analyse these systems that can assist in the development of systems that are responsive to the dynamics of real manufacturing environment. As well, the insights gained from these modelling techniques are needed to achieve the important goal of objectively evaluating the relative performance of alternative control strategies.
Given these requirements, a considerable amount of research has been conducted recently on the analysis of discrete event dynamic systems (DEDS), or man-made systems such as manufacturing systems that are characteristically event-driven and asynchronous in nature. These systems differ from continuous variable dynamic systems (CVDS) of classical control theory (where there is already a strong base of research) primarily in the fact that they cannot be modelled using traditional modelling techniques (i.e., differential equations) that are common and well established for the solution of CVDS problems.
Although, DEDS theory can be considered to be in its infancy when compared to CVDS theory, there have been a number of advances in this area in recent years. Of primary interest to this research are advances concerned with moving discrete-event simulation from an evaluative modelling tool to a generative modelling tool: i.e., although simulation can be used to evaluate the performance of a manufacturing system (e.g., flow time and due date performance) for given parameter values (e.g., station cycle times), it cannot generate new parameter values that lead to improved manufacturing system performance (i.e., simulation does not provide sensitivity information at the end of a single experiment). Recent research into single-run gradient estimation using perturbation analysis, as well as research into using these techniques for stochastic optimisation have increased the power of discrete event simulation as a modelling tool for manufacturing. There have been considerable advances in the development of these analytical techniques, particularly in the area of expanding the classes of systems that can be analysed, yet there is little research that is being conducted on applications of these techniques to specific problems such as manufacturing control.
R.W. Brennan and B. Foroughi, "A control framework to support responsive manufacturing," the International Journal of Agile Management Systems, Vol. 1, No. 3, pp. 159-168, 1999.
R.W. Brennan and P. Rogers, "Stochastic optimization applied to a manufacturing system operation problem," Winter Simulation Conference Proceedings, ed. C. Alexopoulos, K. Kang, W.R. Lilegdon, and D. Goldsman, pp. 857-864, 1995.
Email: brennan@enme.ucalgary.ca Telephone: (403) 220-5798
Last updated: 16 June 2003