A multi‐level approach for multi‐objective optimization in process eco‐design

Abstract This work addresses a central challenge in process eco‐design: how to efficiently integrate environmental and economic criteria—often conflicting—into early design stages while managing the high dimensionality of Life Cycle Assessment (LCA) indicators. To this end, we propose a generic six‐level multi‐objective methodology combining process simulation, LCA, a modified NSGA‐II genetic algorithm, and multiple criteria decision‐making (MCDM) models (ELECTRE, PROMETHEE, M‐TOPSIS). The approach is applied to the hydrodealkylation of toluene (HDA) for benzene production. Starting from a set of five criteria (four environmental end‐points and economic performance), later expanded to 14 environmental mid‐points, the method identifies redundant indicators and reduces the optimization to the most influential objectives. Results show that a bi‐objective optimization (economic performance vs. human‐health impacts) is sufficient to represent trade‐offs previously captured by the full 14‐indicator space. Quantitatively, the optimized solutions achieve 12%–18% reductions in environmental end‐point impacts relative to single‐objective economic optimization, without increasing operating costs. Mid‐point results indicate that ozone layer depletion is the dominant contributor among human‐health categories, guiding the formulation of a refined optimization problem. MCDM techniques consistently identify similar compromise solutions, confirming the robustness of the reduced‐criteria strategy. Overall, this framework demonstrates that combining LCA, multi‐objective optimization, and MCDM into a structured multi‐level workflow enables engineers to reduce problem complexity while achieving environmentally improved designs that remain economically viable.