Optimal synthesis of multiproduct batch plants with cyclic scheduling and inventory considerations
journal contributionposted on 01.01.1992 by V. T. Voudouris, Ignacio E. Grossmann, Carnegie Mellon University.Engineering Design Research Center.
Any type of content formally published in an academic journal, usually following a peer-review process.
Abstract: "This paper addresses the problem of determining the optimal configuration and cyclic operation of batch plants in which all the products require the same processing sequence. In particular, the problem can be stated as follows. Given are demands of a number of products, as well as technical information on the processing tasks (size factors, processing times, clean-up times) which are not restricted to a zero-wait policy. Given are also cost data for investment and product inventories, a list of candidate equipment and a list of candidate storage vessels with standard sizes. The problem then consists in determining the following items: number, type and size of equipment, as well as their allocation to one or multiple tasks and possible parallel operation; location and size of intermediate storage vessels; the length of the production cycle including the sequence of production of the products; levels of product inventories. The objective is to maximize the net present value. The major complication of this design problem lies in the many trade-offs that are involved, as for instance the merging of tasks versus its impact on the schedule, and length of production cycle versus inventory levels. By using a novel representation for cyclic schedules and exact linearization schemes, it is shown that this problem can for [sic] formulated as a mixed- integer linear programming problem, and solved rigorously to global optimality. An efficient computational scheme is proposed for this purpose. Compared to the previous work by Birewar and Grossmann (1990), the proposed model provides a significant extension of the scope of the operational problem, while at the same time yielding an optimization problem that does not involve nonlinearities. Several example problems are presented to illustrate the capability of this method."