MIXED-INTEGER LINEAR-PROGRAMMING REFORMULATIONS FOR BATCH PROCESS DESIGN WITH DISCRETE EQUIPMENT SIZES

The objective of this paper is to show that many nonlinear models for batch design which are based on the assumption of continuous sizes can be reformulated as MILP problems when sizes are restricted to discrete values. Problems considered include multiproduct plants operating with single product and mixed product campaigns, and multipurpose plants with single and multiple production routes. It is shown that, by exploiting the structure of the proposed MILP models, solutions can be obtained with modest computational effort. In addition, as opposed to the use of rounding schemes for continuous models, global optimum solutions are guaranteed.