Short term scheduling in multipurpose batch chemical plants is specially complicated by the fact that plant capacity is not well defined due to the multipurpose nature of equipment units, allowing different routes. In order to reduce problem dimension, as well as allowing to analyze problem feasibility prior to spending time solving a large scheduling problem, a planning phase is proposed. At this level the objective is to determine a set of processing time windows for the batches necessary to fulfill final products' demand. The time windows allow analyzing equipment units load and plan feasibility in terms of satisfying final products due dates. To accomplish those tools from the area of Constrained Based Search are used. Feasibility analysis is done for a fixed assignment of tasks to equipment units. Through the analysis of equipment units load, the user can modify system's inputs, as raw material delivery plan, assignment task/equipment and product's demands, in order to obtain an acceptable final plan. This planning frame is especially well suited to reduce the dimension of the scheduling problem, as it is shown for a MILP formulation and for a Simulated Annealing application.
multipurpose batch plants; short term planning and scheduling; constraint propagation
