Metaheuristic methods and especially genetic algorithms are the most popular simulation
optimization methods. Genetic algorithms are perhaps the most mature metaheuristic methods
for simulation optimization in inventory management; they have been applied to dierent types
of inventory management problems and often are compared with other sim-opt methods in
terms of their performance. Metamodel-based methods appear less popular than metaheuristic
or hybrid methods though, such that they appear mainly in methodology focused texts. Yet they
also oer powerful means for solving robust sim-opt and stochastically constrained inventory
problems. Applying metamodel-based methods (especially kriging, as stochastic kriging opens
new opportunities to account for simulation noise) to practical inventory problems can be a
promising area for research.