<p>Task offloading in IIoT is a challenging problem with multiple objectives to be handled optimistically since there are conflicting goals between minimizing the task execution delay and the energy savings, constrained by decisions. For this purpose, in this paper, a Multi-objective Ranking-based Discrete Firefly Algorithm with Decomposition (MO-RDFA/D) is proposed for task offloading optimization to mobile edge computing system. The proposed algorithm combines continuous firefly search with a new ranking-based discretization approach to achieve feasible task-to-server mappings, and it uses Tchebycheff-based decomposition and Pareto dominance to handle multiple objectives in an efficient manner. Simulation results are carried out by considering different densities of users (i.e., 200–1000 user) and heterogeneous ETC scenario. Extensive experiments show that MO-RDFA/D can achieve better results than the current competitive evolutionary, heuristic and learning-based methods. More precisely, compared to some state-of-the-art schemes, the proposed scheme decreases average energy consumption up to 10.3%, reduces average task delay by up to 16.9% and prolongs network lifetime at most 6.7%. The multi-objective performance evaluations of MO-RDFA/D compared with ONVG and SP indexes also indicate that higher Pareto front coverage (26.87 non-dominated solutions at most) and better distribution of solutions are obtained by using MO-RDFA/D (SP being as low as 0.10). These results validate that MO-RDFA/D is an efficient and scalable candidate in dealing with discrete multi-objective task offloading problems for dense IIoT environments.</p>