A comparison of the performance of multi-objective optimization methodologies for solvent design

Abstract In this work, we present a systematic comparison of the performance of five mixed-integer non-linear programming (MINLP) multiobjective optimisation algorithms on a computer-aided solvent design problem. The five methods are designed to address the nonconvexity of the problem, with the aim of generating an accurate and complete approximation of the Pareto front. The approaches includes: a weighted sum approach with simulated annealing (SA), a weighted sum approach with multi level single linkage (MLSL), the sandwich algorithm with SA, the sandwich algorithm with MLSL and the non dominated sorting genetic algorithm-II. These five combinations of optimisation techniques are applied to the design of a solvent for chemical absorption of carbon dioxide (CO2). The results shows that the sandwich algorithm with MLSL can efficiently generate diverse Pareto points leading to a construction of more complete Pareto front.