Design and Analysis of Simulation Experiments

Design and Analysis of Simulation Experiments

Preface

This book is the successor of several other books that I wrote on (roughly) the same topic. My first book consisted of two volumes, and was published in 1974/1975 (and translated into Russian in 1978). Its successor was pub- lished in 1987. In 1992, Willem van Groenendaal and I wrote a more general book on simulation, which included an update of parts of my 1987 book.

So I thought that it was high time to write down all I know about the statistical Design and Analysis of Simulation Experiments, which I abbre- viate to DASE (and pronounce as the girl’s name Daisy). This acronym is inspired by DACE, which stands for Design and Analysis of Computer Experiments; the acronym DACE is popular in deterministic simulation.

In this book, I will focus on those DASE aspects that I have a certain expertise in—I think. Though I focus on DASE for discrete-event simulation (which includes queueing and inventory simulations), I also discuss DASE for deterministic simulation (applied in engineering, physics, etc.).

I discuss both computationally expensive and cheap simulations. I assume that the readers already have a basic knowledge of simulation; e.g., they know concepts such as terminating simulation and steady-state simulation. They should also have a basic understanding of mathematical statistics, including concepts such as distribution functions, averages, and variances.

This book contains more than four hundred references. Yet, I have tried to eliminate older references that are mentioned in more recent references— unless the older reference is the origin of some important idea (so the readers may get a historical perspective). To improve the book’s readability,I try to collect references at the end of paragraphs—as much as seems reasonable.