Design and Analysis of Computer Experiments

Abstract

Robust design can be performed at an early stage when the engineer still has a substantial amount of freedom. The problems that are addressed with virtual experiments include (i) to understand which factors have the biggest influence on the response and (ii) to build a mathematical model. Two very different kinds of experimental designs are used for virtual experiments. One is screening designs; the other type is space filling designs. The type of response surface most widely used in industrial practice is probably kriging. The reason could be that kriging can fit a surface that actually goes through each calculated point. Once the response surface is in place, it is time to pick up the distinction between noise and control parameters in order to find a robust solution. Optimization methods offer a more solid way to proceed and are easily available using commercial software.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 181 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.