Multi-Domain Optimisation Using Computer Experiments for Concurrent Engineering

It is a challenge to optimise several simulators in a multi-objective way in Computer-Aided Engineering (CAE). The technology of replacing each simulator with a fast statistical emulator based on a space-filling computer experiment is introduced. A case study based on a rudimentary model of a car suspension system is conducted in full and is one of a series in a collaborative EU project (No. BE96-3046) which has a particular emphasis on multi-domain problems: mechanical, thermal, electrical etc. In the case study the objectives from three separate simulator responses are combined in different portmanteau criteria; global optimisation is then used to find Pareto solutions.