Surrogate-Assisted Robust Multi-Objective Optimization of a Fluidized-Bed Methane Pyrolysis Reactor Under Technical and Market Uncertainty

Conventional multi-objective optimisation (MOO) of methane pyrolysis reactors assumes model parameters as precisely known, while many crucial factors are highly unknown. This research uses NSGA-II and Monte Carlo sampling to produce and analyse robust Pareto fronts under uncertainty using a surrogate-assisted robust multi-objective optimisation framework. Kriging surrogates trained on 300 Latin hypercube designs with 500 Monte Carlo samples compare mean–variance, worst-case, and chance-constrained formulations.