DESIGN OF SINGLE-PRODUCT CAMPAIGN BATCH PLANTS UNDER DEMAND UNCERTAINTY

A new method is introduced for optimally designing multiproduct batch plants under the single-product campaign (SPC) production mode. Uncert ain future product demands are described with normal probability distr ibutions and more than one processing unit of equal size are allowed p er stage. At the expense of imposing the normality assumption for prod uct demand uncertainty and the SPC production mode, the original two-s tage stochastic optimization problem is transformed into a determinist ic mixed-integer nonlinear programming problem without relying on impl icit or explicit discretization of the uncertain variables. This is ac complished through the explicit solution of the inner problem and the analytical integration overall product demand realizations. This probl em representation and solution strategy result in savings of orders of magnitude over existing methods in computational requirements.