Optimization-based design of reactive distillation columns using a memetic algorithm

This contribution deals with the optimization of the design of reactive distillation columns by using a memetic algorithm (MA) which is a combination of an evolution strategy (ES) and a mathematical programming (MP) solver. The focus of this paper is on the restriction on the number of feed trays that introduces discrete variables into the problem formulation. The optimization of the number and of the location of the feed streams is addressed by the EA which can deal with discontinuous cost functions and integrality constraints. A minimization of the number of feed streams is achieved by adding penalties to the fitness function of the EA. The results of the new approach are compared to the results of the optimization without restriction on and the minimization of the number of feed streams.