Kriging-assisted constrained optimization of single-mixed refrigerant natural gas liquefaction process

This paper presents a surrogate-assisted framework to solve constrained black-box optimization problems, which is applied the optimal single-mixed refrigerant natural gas liquefaction processes design. approach uses kriging models as cheap-to-evaluate, reliable surrogates of process-simulator-dependent functions problem. From prediction and estimated error, probability feasible improvement acquisition function derived optimized search forfeasible improving candidates perform local in penalized simulation. The kriging-assisted global scheme efficient find promising for refinement, it improves previous results that used only by 3.36 5.32%. surrogate model reduces time evaluate from 0.9933 ± 0.0815 3.515e-4 8.353e-5 s. approximation mean squared errors assessed k-fold cross-validation 7.271e-7 1.297e-7 objective 0.3099 0.0856 constraint.