Computer simulations for the optimization of technological processes

This chapter is about experiments for quality improvement and the innovation of products and processes performed by computer simulation. It describes familiar methods for creating surrogate models of simulators (emulators), with particular reference to Kriging interpolation, and some new ways of fitting the models to the simulated data. It also deals with the advantages of computer experiments performed sequentially, and with computer experiments in which some of the random noise factors that affect the output of a process are simulated stochastically. As an example, an application to the integrated parameter and tolerance design of a high-precision space-measuring instrument is illustrated, and other potential applications are also mentioned.