A tutorial review of techniques for simulation optimization

Reviews techniques for optimizing stochastic discrete-event systems via simulation, for both discrete and continuous parameters. For discrete parameters, we focus on the techniques for optimization from a finite set: multiple-comparison procedures and ranking-and-selection procedures. For continuous parameters, we discuss sequential response surface methodology procedures and stochastic approximation gradient-based procedures, and describe gradient estimation based on perturbation analysis, likelihood ratio and frequency domain experimentation. We then discuss two applications: an inventory control problem with a "noisy" constraint and a call option pricing problem in finance.