(356e) Optimal Scheduling of An Industrial Food Manufacturing Facility

Scheduling of batch and semi-continuous processes is tied intimately with the efficiency of operations, utilization of storage, demand satisfaction, and operations uptime. Process scheduling is complicated by interdependencies of equipment networks, complex product recipes, storage limitations, product due dates (material deliveries) and limited utilities. Given the above considerations, the development of mathematical models to optimally schedule processes has received considerable attention within the process systems engineering community.

The main differentiation between the various modeling methodologies for short-term scheduling lies in the representation of time. Discrete-time models subdivide the scheduling horizon into evenly spaced intervals and allow events to occur only at the interval boundaries. A key contribution to discrete-time scheduling models was made by Kondili et al. (1993) with the introduction of the state-task-network (STN) paradigm. Continuous-time models introduce the exact timing of events through the introduction of event points whose location is not specified a priori, but determined through the solution of the optimization problem (Ierapetritou and Floudas, 1998). Excellent reviews of scheduling model formulations are provided in Floudas and Lin (2004) and Mendez et al. (2006).

In this paper, we present an application of a continuous time STN framework to the scheduling of an industrial food manufacturing facility. The facility has a large number of products that are also subject to variation in packaging type and/or size. Sequence-dependent cleaning of process equipment is also required and production runs are restricted to specific product groups. A process-wide equipment cleaning task is also necessary to reduce the risk of bacterial growth and subsequent product contamination. These cleaning events must adhere to a specified minimum frequency as specified through quality and assurance requirements.

The proposed formulation is based on the approach of Maravelias and Grossmann (2004), where particular attention is paid to efficient modeling of the cleaning events. We show first an application to events of fixed timing, and thereafter present extensions that allow events to occur within bounded times and that capture the cleaning requirements of the industrial manufacturing application under consideration. Case studies are presented that compare actual schedules generated by plant management against those generate using the described modeling framework, including scenarios in which rescheduling was required due to events that were unforeseen at the start of the production run. The computational performance of the optimization approach is also discussed. Insights learned in industrial scheduling and planning problems of this type are presented, and future research directions discussed.

References:

C.A. Floudas and X. Lin. Continuous-time versus discrete-time approaches for scheduling of chemical processes: a review. Computers and Chemical Engineering, 28: 2109?2129, 2004.

M.G. Ierapetritou and C.A. Floudas. Effective continuous-time formulation for short-term scheduling. 1. Multipurpose batch processes. Industrial & Engineering Chemistry Research, 37:4341?4359, 1998.

C.T. Maravelias and I.E. Grossmann. A general continuous state task net-work formulation for short term scheduling of multipurpose batch plants with due dates. In Proceedings of the 8th International Symposium of Process Systems Engineering (Editors: Chen, B.; Westerberg, A.W.), 274-279, Kunming, China, 2004. Computer-aided Chemical Engineering, Vol. 15, Elsevier.

C. A. Mendez, J. Cerda, I.E. Grossmann, I. Harjunkoski, and M. Fahl. State-of-the-art review of optimization methods for short-term scheduling of batch processes. Computers and Chemical Engineering, 30:913 ? 946, 2006.

E. Kondili, C.C. Pantelides and R.W.H. Sargent. A general algorithm for short-term scheduling of batch operations-I. MILP formulation. Computers and Chemical Engineering, 17(2):211 ? 227, 1993.