A new multi-objective optimization algorithm for separation processes

This paper proposes a new algorithm, named guided population adaptive genetic algorithm (GAGA), to solve optimization problem of the complex separation processes. In GAGA, the strategies of adaptive crossover and mutation, leader selection, boundary random walk, neighborhood guidance and spiral updating position are introduced to enhance the guidance of population evolution. The capability of GAGA is investigated by 8 multiobjective benchmark problems. The results are compared with four well-known multi-objective optimization algorithms. The average values of inverted generational distance (IGD) and generational distance (GD) are 0.1626 and 0.5363, respectively, which is proved to be a robust and reliable model. Moreover, GAGA is validated through methacrylic acid (MAA) separation process with multi-cycle flows. The optimization efficiency of GAGA has accelerated by 2 times compared with non-dominated sorting genetic algorithm-II (NSGA-II), with results of 4.7 % reduction in total annual cost (TAC) and 4.3 % reduction in global energy consumption (GEC).