Multi-Product Scheduling in a Chemical Plant

Details are given of a chemical plant consisting of four machines, each capable of producing one of several products at a time. Five products are to be produced, each subject to a normal demand distribution with known parameters. The rates of production, costs, profit margins, setup times and machine capacities are given.

The study compares two production scheduling methods, the first (Method I) is based on linear programming (two variants of this method are considered) and the second (Method II) on a multi-product batch scheduling approach.

A simulation of the plant for a variety of demand conditions and for several safety-stock policies was carried out. The results suggest that Method I yields somewhat better plant performance for the case when average demand is below plant capacity, but neither method showed superiority under all conditions, and in many cases the difference between the results for the two methods was only marginal.