This book covers the practice of running a computer experiment, for which the author is uniquely qualified: not only has he contributed to the scientific foundations of the topic, but, as both an academic and a consultant, he has designed and run a wide variety of computer experiments, in environmental science, physics, and industrial applications. The main topics for computer experiments are uncertainty propagation (how does uncertainty in the simulator inputs affect uncertainty in the outputs), sensitivity analysis (which inputs matter for which outputs), tuning and calibration (learning about the inputs from measurements on the system), prediction (of the system, possible under a variety of interventions), and optimization. Throughout the book, the focus is on practical issues, using consistent notation and definitions, there is substantial practical guidance on running a computer experiment, including organization and communication. Code snippets, in R, are used to illustrate algorithms, and the layout of script files.

This practical guide will be essential reading for any scientist or engineer who performs computer simulations, enabling researchers to develop and perform more robust experiments more easily and understand how their model functions.

Key Features:

• Written by a practitioner, for other practitioners

• Covers only the useful theory, with focus on practical suggestions

• Many pitfalls illustrated with examples

• Short, self-contained chapters for easy use

• Consistent use of definitions and notation for clarity and communication