Evolutionary Design Method for Multipurpose Batch Plants on the Basis of Cyclic Production.

Abstract

Designing of multipurpose batch plants is a problem of optimizing the plant configuration and the equipment sizes under the constraint of the production plan. However, these decision variables are strongly interactive with each other through scheduling, so that simultaneous optimization is impossible.In this study, the design problem of multipurpose batch plants is considered on the basis of cyclic production, and the evolutionary method is adopted from the analysis of this design problem. The plant configuration is decided as the evolutionary variable, and two IP (Integer Programming) models are developed to generate the initial and/or neighboring plant configurations. The initial plant configuration with minimum number of equipment units is obtained through these IP models, and the equipment sizes under this initial plant configuration are optimized through cyclic scheduling. The neighboring plant design can be generated repeatedly through the developed IP model and the cyclic scheduling from this initial one, and an evolutionary design method for the multipurpose batch plants is developed. The effectiveness of the developed method is demonstrated through an example design problem.