On the use of filters to facilitate the post-optimal analysis of the Pareto solutions in multi-objective optimization

• We propose using Pareto filters to automatically select the most promising solutions of a multi-objective optimization problem.
• The solutions kept by the filters are those showing better overall performance.
• It is crucial to assess not only the number but also the quality of the solutions kept.
• Our strategy is applied to the MOO of a reverse osmosis desalination plant.

Multi-objective optimization (MOO) has emerged recently as a useful technique in the design and planning of engineering systems because it allows identifying alternatives leading to significant environmental savings. MOO models typically contain an infinite number of Pareto solutions, from which decision-makers should choose the best one according to their preferences. An approach is here presented that identifies and retains for further inspection a reduced set of Pareto solutions showing better overall performance. The capabilities of our approach are illustrated through its application to the design of reverse osmosis desalination plants considering simultaneously the unitary production cost and a set of environmental impacts in several damage categories. Our method reduces significantly the number of Pareto points, thereby facilitating the decision-making process in MOO.