Accuracy of neural network approximators in simulation-optimization

Heuristic search techniques are highly flexible but computationally intensi ve optimization methods that require hundreds, sometimes thousands, of eval uations of the objective function to reach termination criteria in common w ater resources optimization applications. One way to make these techniques more tractable when the objective function depends on a time-consuming flow and transport model is to employ an empirical approximation of the model. The current study examines the impact of employing artificial neural networ ks (ANNs) and linear approximators (LAs) on the quality and quantity of sol utions obtained from simulated annealing-driven searches on two different g round-water remediation problems. The quality of results obtained when ANNs served as substitutes for the full model was consistently comparable to th at of results obtained when the full model itself was called in the course of the search. The effect on quality of results of substituting an LA for t he full model was more variable.