Integration of production planning and scheduling in the chemical industry

Christos Maravelias, Chemical Engineering, University of Madison-Wisconsin

26 March 2009 at 10:30

Location: JHE 326H

We present an overview of supply chain management and the associated planning problems. We then formally define the problem of medium-term production planning and explain why integration with short-term scheduling is often necessary. We discuss three approaches to address this integrated planning-scheduling problem. First, we present a novel formulation for the multi-item capacitated lot-sizing problem. The formulation accounts simultaneously for a) sequence-independent setups that carry over from previous periods, b) setups that cross over planning period boundaries, and c) setups that are longer than one period. It is extended to model time periods of non-uniform length, parallel units, families of products, backlogged demand, and lost sales. Second, we present how detailed mixed-integer programming (MIP) scheduling models can be used off-line to obtain a convex approximation of feasible production levels and a convex underestimation of total production cost as a function of production levels. We also show how this idea can be extended to model non-convex functions. Third, we study the properties of the polyhedron defined by the constraints of a discrete-time MIP production planning formulation. We prove that a submatrix of this problem has the property of k-regularity, thus implying that these problems can be solved very effectively. Finally, we present how these methods can be used to address large-scale integrated production
planning/scheduling problems.

Return to list of all seminars