Non-dominated Sorting Genetic Algorithm II (NSGA-II) remains one of the most widely adopted metaheuristics for solving multi-objective optimization problems; however, its conventional usage is often limited to default parameter settings and moderate-scale problem instances, which restricts its effectiveness in large-scale and high-dimensional scenarios. This tutorial presents a comprehensive and practical guide to scaling NSGA-II for large-scale optimization, emphasizing the critical role of algorithm configuration, parameter tuning, and implementation design. We present empirical evidence demonstrating that appropriately configured NSGA-II variants can effectively address both benchmark and real-world problems involving extremely high-dimensional decision spaces. The tutorial further discusses computational considerations, including parallelization strategies and memory efficiency, providing actionable insights for practitioners aiming to deploy NSGA-II in high-performance computing environments. Overall, this tutorial highlights that careful algorithmic configuration and scalable implementation are key to unlocking the full potential of NSGA-II in large-scale multi-objective optimization.
