Our civilization employs increasingly complex human-made systems, such as large-scale electric power grids, air traffic control systems, manufacturing plants and supply chains, the Internet and other communication networks. Performance evaluation of these systems is accomplished by using simulation models rather than experiment with the real systems. However, these systems operate and evolve in time via human-made rules of operation which are difficult to describe and capture by succinct mathematical models. And while simulation models are often used for design validation and other purposes, computational constraints and the changing nature of the problem domain make them unsuitable for optimization purposes.

If we accept the need for search based methods as a complement to the more established analytical techniques, then quickly narrowing the search for optimum performance (ordinal optimization) is more important than accurately estimating the values of system performance during the process of optimization (cardinal optimization). The purpose of this book is to address the difficulties of simulation optimization problems — the optimization of complex systems via simulation models or other computation-intensive models involving possible stochastic effects and discrete choices. This book will establish the distinct advantages of the "softer" ordinal approach for search-based type problems, analyze some of its general properties, and show the many orders of magnitude improvement in computational efficiency that is possible. As such, the book is complementary to existing optimization literature. The tools described here do not replace but can be used separately or in conjunction with other methodological tools of optimization.