All simulation models are descriptive as opposed to prescriptive. Therefore, a simulation analyst is required to analyze and interpret the simulation model’s output data. For complex models, however, this analysis is extremely difficult. The author provides an excellent description of the metamodeling technique to help simulation analysts overcome this difficulty.
A metamodel is a mathematical model of a simulation model’s input-output relations. It is used to further abstract the real system’s input-output transformation and help the analyst better understand the input-output relations of the simulation model and the real system.
This is a well-written, short book with six chapters, an appendix, and a bibliography. The book begins with an introduction to simulation modeling and metamodeling. The simulation metamodel and its role in different levels of abstraction are described in chapter 2. Chapter 3 provides a high-level review of the statistical analysis of simulation output data and experimental design. The development of multivariate simulation metamodels and the assessment of their credibility are presented, with an example, in chapter 4. Chapter 5 surveys the current research in and applications of simulation metamodeling. The book concludes with chapter 6, which provides two additional examples of metamodeling.
This is an excellent book on simulation metamodeling. It can be used to supplement a general-purpose textbook in a simulation course. It should be read by any researcher or practitioner interested in the statistical analysis of simulation models.