Efficient optimization algorithms for zero wait scheduling of multiproduct batch plants

Abstract: "This paper deals with the scheduling of multiproduct batch plants consisting of one unit per stage and involving the production of large number of batches of relatively few products (e. g. 10-50) with the Zero-Wait policy. An aggregation procedure based on a compact linear programming model is proposed for determining the minimum cycle time and the family of schedules with the optimum cycle time. It is shown that schedules can be easily derived from the LP solution using an aggregated graph representation. It is also shown that using this representation an approximate MILP model that yields very good near optimal solutions can be derived for the makespan minimization problem and for which very tight bounds can be established. The insights gained with this approximate method are used to formulate a rigorous MILP model which can be solved as an LP in most cases. The effectiveness of the proposed procedures is illustrated with several example problems.