Application methods for a niche genetic algorithm for design of reactive distillation processes

A purpose of this chapter is to investigate application methods for a niche GA to design of Reactive Distillation (RD) process involving the production and separation of ethyl acetate. First, optimization system based on the Multi- Niche Crowding – Genetic Algorithm (MNC-GA) is demonstrated to be effective in searching preferable designs of RD process using multiple feeds from each reactant. It is seen that the MNC-GA allows the search to yield various design solutions without causing remarkable performance degradation for searching the best design. Thus, authors investigate application methods for utilizing the various design solutions from the viewpoint of analysis of operability of searched process. It is shown that the multiple preferable designs obtained using the MNC-GA are useful for searching more preferable designs by steady-state process simulation and for sensitivity analysis which provides us with insights to the dynamics in RD column. In addition, intermediate simulation results, which are obtained through all the generations, are demonstrated to produce informative distributed plots for the sensitivity analysis, as compared to the normalGA where the tournament-based selection method is used.